Download Bugging the cell wall of bacteria

Document related concepts

Cell encapsulation wikipedia , lookup

Cytosol wikipedia , lookup

Cell culture wikipedia , lookup

Cellular differentiation wikipedia , lookup

Flagellum wikipedia , lookup

Cell cycle wikipedia , lookup

Theories of general anaesthetic action wikipedia , lookup

Extracellular matrix wikipedia , lookup

Cell growth wikipedia , lookup

Organ-on-a-chip wikipedia , lookup

Amitosis wikipedia , lookup

Mitosis wikipedia , lookup

Lipid bilayer wikipedia , lookup

Lipopolysaccharide wikipedia , lookup

Signal transduction wikipedia , lookup

Model lipid bilayer wikipedia , lookup

Cytokinesis wikipedia , lookup

Cell wall wikipedia , lookup

Lipid raft wikipedia , lookup

Cell membrane wikipedia , lookup

Endomembrane system wikipedia , lookup

List of types of proteins wikipedia , lookup

Transcript
Bugging the cell wall of bacteria
Novel insights into peptidoglycan
biosynthesis and its inhibition
Afluisteren van de bacteriële celwand
Nieuwe inzichten in de biosynthese van peptidoglycaan en
inhibitie daarvan
(met een samenvatting in het Nederlands)
Proefschrift
ter verkrijging van de graad van doctor aan de Universiteit Utrecht
op gezag van de rector magnificus, prof. dr. J.C. Stoof,
ingevolge het besluit van het college voor promoties in het openbaar te
verdedigen op dinsdag 15 juni 2010 des middags te 12.45 uur
door
Nick Kenji Olrichs
geboren op 5 januari 1980 te Delft
Promotor: Prof. dr. B. de Kruijff
Co-promotor: Dr. E. J. Breukink
Het in dit proefschrift beschreven onderzoek werd (mede) mogelijk gemaakt met financiële
steun van EUR-INTAFAR en the Netherlands Proteomics Centre.
Drukwerk: Ridderprint b.v.
ISBN 978-90-393-5350-9
Contents
Chapter 1
General introduction
5
Chapter 2
A novel in vivo cell wall labeling approach sheds
new light on peptidoglycan synthesis in Escherichia coli.
33
Chapter 3
Development of a novel proteomics approach for the
59
identification of proteins interacting in vivo with the bacterial
cell wall.
Chapter 4
Probing the Lipid II binding site of Escherichia coli
penicillin-binding protein 1b.
81
Chapter 5
Mechanism of action of the transglycosylation inhibitor 5b
97
Chapter 6:
Summarizing discussion
113
Nederlandse samenvatting
123
Dankwoord
129
List of publications
132
Curriculum Vitae
133
Abbreviations
135
Chapter 1
General introduction
Partially based on: van Dam, V., Olrichs, N. K., Breukink, E., Specific labeling of
peptidoglycan precursors as a tool for bacterial cell wall studies. Chembiochem. 2009
10(4):617-24
Background
Bacteria are unicellular organisms, which are thought to have been among the first living
organisms on this planet. They have survived for millions of years by adapting to their
ever-changing surroundings. Life as it is for humans would be impossible without the
presence of bacteria, as they perform essential processes such as fixation of nitrogen from
the atmosphere and the decomposition of organic waste. Furthermore, bacteria perform
many beneficial functions in the human body, e.g. the biosynthesis of several vitamins,
digestion of food and development of the immune system. Unfortunately, bacteria are also
responsible for a great number of infectious and sometimes life-threatening diseases.
During the last century, researchers discovered compounds that could kill or inhibit the
growth of bacteria, termed antibiotics. The first natural antibiotic to be discovered and
practically used, penicillin, became one of the most efficacious life-saving drugs in the
world. With the discovery of other antibiotics, a vast pharmaceutical industry was
generated, which would overcome several of mankind's most ancient afflictions, including
tuberculosis, gangrene and syphilis. Most antibiotics used today are synthetically modified
versions of natural compounds, as misuse of antibiotics contributed to the development of
bacteria that have become resistant to the first pool of antibiotics. The last few decades saw
an alarming rise hereof and the emergence of multiple drug resistance, including strains
resistant to vancomycin, the antibiotic of last resort, and the infamous methicillin-resistant
Staphylococcus aureus (MRSA) that is resistant to the large group of β-lactam antibiotics.
This has turned the development of new antimicrobial compounds into a crucial necessity.
Despite the large number of antibiotics, their primary targets can be categorized into a few
groups: DNA replication, protein biosynthesis, cell membrane integrity, production of
essential metabolites and biosynthesis of the cell wall. The bacterial cell wall forms an
invaluable target for antibiotics as it is essential for the viability of the cell and its location
at the cell’s exterior gives it a relatively high accessibility. Well-known antibiotics such as
penicillin and vancomycin kill bacteria by disrupting cell wall synthesis at distinct stages,
thereby weakening the cell wall and ultimately causing lysis. Yet many aspects concerning
the biochemical processes involving the cell wall remain unclear, which likely conceals
potential targets for new antibiotics. It is therefore important to deepen our understanding
of cell wall physiology, which requires the development of additional strategies for cell
wall analysis.
6
Bacterial cell wall: Peptidoglycan structure and biosynthesis
The cell wall surrounding bacterial cells is an absolute vital structure for cell survival. The
main function of the bacterial cell wall is to provide the cell with strength required to
withstand the high internal osmotic pressure. The only class of bacteria devoid of a cell
wall, the mycoplasmas, are therefore osmotically fragile. They have a parasitic lifestyle and
rely on their host cells for a stable environment [1]. Other functions of the cell wall are to
serve as a scaffold for the anchoring of other components of the cell envelope including
proteins and polysaccharides and to contribute to the preservation of a defined cell shape.
The main constituent of the cell wall is peptidoglycan (murein), a giant macromolecule
consisting of a multilayered network of linear glycan strands interlinked with each other via
short peptide bridges. Throughout the cell cycle, peptidoglycan is continuously remodeled
to allow cell growth and division. In Gram-positive bacteria, a dense peptidoglycan cell
wall up to 40 layers thick (20-80 nm) is present at the cell’s exterior. Other major
constituents of the Gram-positive cell wall are teichoic acids, lipoteichoic acids and cell
wall proteins bound either covalently or non-covalently to the peptidoglycan (Fig. 1)[2, 3].
Gram-negative bacteria only have a thin single peptidoglycan layer (~4 nm) located in the
periplasmic space between the cytoplasmic (inner) membrane and an additional protective
outer membrane [2]. The inner leaflet of this highly asymmetric outer membrane consists of
phospholipids, whereas the outer leaflet is composed of lipopolysaccharides. An abundant
class of transmembrane channel proteins in the outer membrane are the porins, which have
a characteristic β-barrel structure and are non-covalently associated with peptidoglycan. A
special group of proteins, the lipoproteins, are covalently bound to peptidoglycan, such as
Braun lipoprotein (Lpp) and peptidoglycan-associated lipoprotein (Pal), which anchor the
outer membrane to the peptidoglycan sacculus (Fig. 1)[3-5].
The glycan strands of peptidoglycan are composed of alternating β-1,4-linked Nacetylglucosamine (GlcNAc) and N-acetylmuramic acid (MurNAc) units (Fig. 2). The
aminosugars occasionally display minor modifications including O- and N-acetylation,
which probably occur at late stages of cell wall maturation [6]. The lactoyl group of the
MurNAc residue is substituted by a peptide stem, the composition of which is most often Lalanine-γ-D-glutamate-meso-2,6-diaminopimelic acid (or L-lysine)-D-alanine-D-alanine in
nascent peptidoglycan, whereas some terminal D-alanine residues are lost in the mature
macromolecule as a result of peptide crosslinking (Fig. 2). The composition of the peptide
shows great variation among the various bacterial species [6-8]. Diaminopimelic acid
7
Porin
Lipopolysaccharide
Teichoic acid
Lipoteichoic acid
Lipoprotein
Peptidoglycan
Membrane of a cross-section of the Gram-negative (left) and Gram-positive
Figure 1. Schematic representation
(right) cell envelope.
(DAP) is present in virtually all E. coli and other Gram-negative bacteria but can also be
found in peptidoglycan of Gram-positives like some lactobacilli, clostridia, bacilli,
corynebacteria and propionibacteria [6]. By contrast, L-lysine is rarely observed in Gramnegative species [9]. Other variants at this and other amino acid positions are less
frequently encountered. Cross-linking of the glycan strands usually occurs between D-Ala
at position 4 and the diamino acid at position 3, either directly or via a short peptide bridge
[6, 10]. These cross-links provide the cell wall its rigidity and strength. Chemical features
characteristic for peptidoglycan are the presence of the unusual amino sugar MurNAc, Damino acids and γ-bonded D-glutamate.
The biosynthesis of peptidoglycan is a complex process, which involves numerous steps in
the cytoplasm and the inner and outer leaflets of the cytoplasmic membrane. Synthesis
commences in the cytoplasm (Fig. 3A) where the nucleotide precursors uridine-5’diphosphate-GlcNAc (UDP-GlcNAc) and UDP-MurNAc-pentapeptide are synthesized. The
first reaction is the transfer of enolpyruvate to the 3 position of UDP-GlcNAc by MurA.
The subsequent step is catalyzed by MurB, which reduces the enolpyruvate moiety to Dlactate; yielding UDP-N-acetylmuramic acid. The assembly of the pentapeptide onto UDP-
8
GlcNAc
CH2OH
O
MurNAc
CH2OH
O
HO
O
NH
O
C
CH3
L-Ala
O
NH
HCCH3
C O
O
O
HN
C O
CH3
H3C
CH
O
C
NH
HO
D-Glu
C
C
H
O
CH2
H2C
C
HN
L-Lys/(m-DAP)
O
C
H3C
C
H
H
C
O
(CH2)4
NH
D-Ala
HN
D-Ala
HO
H
C
NH2
C
O
OH
C
O
H
C
CH3
C
O
Figure 2. Schematic of peptidoglycan structure consisting of linear glycan strands interlinked via
peptide side chains (left). Chemical structure of a peptidoglycan monomer unit composed of two
aminosugars, N-acetylglucosamine (GlcNAc) and N-acetylmuramic acid (MurNAc), and a
pentapeptide most often containing L-alanine, D-glutamic acid, L-lysine (or meso-diaminopimelic
acid) and two D-alanines. Peptides from adjacent units are either directly linked or via additional
short peptide bridges.
MurNAc is carried out in sequential steps by MurC, MurD, MurE and MurF. In E. coli,
MurC catalyzes the addition of L-alanine to the lactoyl moiety of UDP-MurNAc. MurD
subsequently adds D-glutamic acid, MurE couples meso-diaminopimelic acid at the γcarboxyl group of glutamic acid and finally MurF attaches the preformed dipeptide Dalanine-D-alanine [11-15] (Fig. 3A).
The transferase MraY, an integral membrane protein, catalyzes the transfer of the phosphoMurNAc-pentapeptide moiety to a membrane acceptor undecaprenyl-phosphate, thereby
forming Lipid I. Thereafter, the GlcNAc moiety is added to Lipid I by the transferase MurG
generating Lipid II (Figure 3B). Thus, Lipid II contains a complete monomer unit of the
peptidoglycan layer of E. coli, GlcNAc-MurNAc-pentapeptide [16, 17]. The mechanism by
which Lipid II is transported across the cytoplasmic membrane has long remained elusive.
Recent studies have revealed that the integral membrane protein FtsW facilitates the
translocation of Lipid II across model and bacterial membranes [18]. FtsW is essential for
9
A
MurA
G
UDP
MurB
M
MurC
UDP
L-Ala
M
MurD
UDP
M
D-Glu
MurE
UDP
L-Lys/
m-DAP
M
MurF
UDP
M
UDP
D-AlaD-Ala
B
GM G M G M GM
GM G M G M GM
PBP’s
G
M
MraY
Lipid II
Lipid I
M
M
MurG
M
G
UDP
G
UDP
Figure 3. (A) Cytoplasmic steps of the peptidoglycan biosynthesis pathway. UDP-GlcNAc is
converted into UDP-MurNAc via MurA and MurB. The amino acids L-alanine, D-glutamic acid and
either L-lysine or meso-diaminopimelic acid are sequentially added by MurC, MurD and MurE,
respectively. The dipeptide D-alanyl-D-alanine is attached by MurF forming UDP-MurNAcpentapeptide. (B) Membrane-bound steps of the peptidoglycan biosynthesis pathway. On the
cytoplasmic side of the membrane, UDP-MurNAc-pentapeptide is linked to undecaprenylphosphate
by MraY, resulting in the formation of Lipid I. GlcNAc is added by MurG, yielding Lipid II, which is
translocated to the periplasmic side of the membrane by FtsW (and most likely RodA during cell
elongation). Polymerization of Lipid II and incorporation into the existing cell wall is catalyzed by the
penicillin-binding proteins (PBPs). Undecaprenylphosphate is recycled following dephosphorylation
of released undecaprenylpyrophosphate.
10
cell division and its structural homologues, RodA and SpoVE, are essential for cell
elongation and sporulation, respectively [19]. It would therefore be logical that these
enzymes are involved in Lipid II translocation during the respective stages of cell growth.
At the exterior of the membrane, Lipid II is utilized as the substrate for polymerization
reactions resulting in the formation of nascent peptidoglycan and its subsequent
incorporation into the preexisting cell wall (Figure 3B). Upon release of the peptidoglycan
monomer unit, free undecaprenyl pyrophosphate is generated, which is dephosphorylated
and transported back to the inner side of the cytoplasmic membrane by a yet unknown
mechanism (Figure 3B). In E. coli, several undecaprenyl pyrophosphate phosphatases have
been identified [20-22]. Evidence has shown that phosphatase activity of LpxT is coupled
to modification of LPS at the periplasmic side [23, 24], however, the location of activity of
the other phosphatases is not clear. At the cytoplasmic side, undecaprenyl phosphate is
again available for the biosynthesis of peptidoglycan and a variety of other cell-wall
polysaccharides.
The polymerization of peptidoglycan involves enzymes exerting two major types of
activities: transglycosylases (TG) and transpeptidases (TP). The transglycosylases catalyze
the elongation of the linear glycan strands and will be discussed in more detail in the next
section. Transpeptidases catalyze the formation of peptide cross-links, between nascent
and/or between nascent and existing glycan strands, thereby enabling the insertion of new
material. Penicillin binding proteins (PBPs) are responsible for the majority of
peptidoglycan synthesized during these final stages (Figure 3B). As their name suggests,
these enzymes are characterized by their ability to bind penicillin and other β-lactam
antibiotics, which causes the inhibition of transpeptidase activity. The PBPs are generally
divided into two classes: the high-molecular weight (HMW) PBPs and low-molecular
weight (LMW) PBPs. All HMW PBPs exhibit transpeptidase activity, with the bifunctional
enzymes additionally catalyzing transglycosylation. E. coli possesses five HMW PBPs of
which PBP1a and PBP1b are the major bifunctional PBPs [25-27]. The functions of some
of the LMW PBPs will be discussed in the section Peptidoglycan breakdown and recycling.
Lipid II as a target for antibiotics
The final membrane-anchored peptidoglycan precursor, Lipid II, plays an essential role in
cell wall biosynthesis. Furthermore, the conservation of a functional biosynthesis pathway
11
for Lipid II in wall-less species indicate its involvement in other cellular processes, such as
cell division [28]. Lipid II is a target for several classes of antibiotics, such as the
glycopeptides and the lantibiotics. They share the common action of preventing
incorporation of Lipid II, thereby weakening the cell wall and eventually killing the cells.
The best known member of the glycopeptide class of antibiotics, vancomycin, was used for
decades as last resort antibiotic. It binds to the terminal D-Ala-D-Ala sequence of Lipid II
and nascent peptidoglycan, thereby sterically obstructing an interaction with peptidoglycan
synthesizing enzymes. Modifications at the peptide moiety of Lipid II, such as the presence
of D-lactate or D-serine instead of D-alanine at the C-terminus, have rendered enterococcal
pathogens resistant to vancomycin [29]. The rapid emergence of bacterial resistance has put
the focus on the therapeutic potential of other Lipid II targeting compounds. Glycopeptide
analogues containing an additional hydrophobic moiety, such as telavancin and oritavancin,
have shown great promise in overcoming vancomycin resistance [30]. Another highly
active antibacterial compound, ramoplanin, was shown to bind Lipid II in a manner distinct
from vancomycin, resulting in the formation of fibrils [31]. The lantibiotic nisin kills
bacteria through so-called targeted pore formation. By the analysis of interactions of nisin
with Lipid II, labeled with pyrene at its lysine, Lipid II was shown to be an integral part of
the nisin pore [32, 33]. The size, stoichiometry, and stability of the nisin-Lipid II pore
complex were elucidated by means of this technology and the extracellular pyrophosphate
of Lipid II was shown to be essential for initial target engagement [34, 35]. This dual mode
of action distinguishes nisin from other pore-forming antimicrobial peptides and gives it
remarkably potent antimicrobial activity. Moreover, other lantibiotics, such as mutacin
1140, were shown to sequester Lipid II into patches without the formation of pores [36].
Most antimicrobial peptides, including nisin, are highly positively charged. The bacterial
cell membrane, mainly composed of the phospholipids phosphatidylethanolamine (70–
80%) and phosphatidylglycerol, has a net negative charge, which facilitates the binding of
the peptides. Resistance to antimicrobial peptide action is developed through changes in the
electrostatic properties of the membrane. A frequently encountered mechanism is the
modification of anionic lipids by the attachment of amino acids, such as lysine, arginine
and alanine. The resulting cationic lipids induce electrostatic repulsion, thereby diminishing
activity of the peptides. Few examples of nisin-resistant strains are known, but its potent
and broad-range activity against Gram-positive bacteria, including multi-drug resistant
strains, make nisin a promising candidate for antibiotic therapy. Most of the described
12
compounds are currently in clinical use or in late-stage clinical trials, highlighting the
importance of Lipid II as a target for antibiotics.
Transglycosylases
The majority of transglycosylation activity is performed by bifunctional PBP’s but is also
found in monofunctional transglycosylases [37]. For long, the potential of the
transglycosylation reaction as a target for the development of novel antibiotics was
recognized, but it could not be fully exploited due to the poorly understood mechanism.
Several classes of compounds are known that inhibit transglycosylase activity by binding to
the common substrate Lipid II. These include the glycopeptides, such as vancomycin, and
the lantibiotics, of which nisin has been the most thoroughly analyzed [32, 38]. The only
well-studied transglycosylase inhibitor to date is moenomycin, a glycolipid produced by
Streptomyces strains which is highly active against a broad range of Gram-positive bacteria.
Moenomycin was discovered some 40 years ago and was touted as a promising new class
of antibiotic, but turned out to be ineffective in humans due to poor pharmacokinetic
properties. It has been extensively used as growth promoter for animals and, somewhat
surprisingly, has not been found to readily induce resistance. With the ominous emergence
of multi-drug resistance the focus has returned to targeting of the transglycosylases [39].
The recent elucidation of several transglycosylase crystal structures, including complexes
with moenomycin, has greatly improved our understanding of the catalytic mechanism and
inhibitory interactions [40-43]. The current model for transglycosylation has the Lipid II
substrate acting as donor and the growing chain as acceptor in the extended binding cleft,
whereas
moenomycin
inhibits
transglycosylation
by
mimicking
the
Lipid
IV
(polymerization product of two Lipid II units) portion of the growing chain in the active site
(Fig. 4). By means of the currently available information, the binding site of the donor
substrate Lipid II can be estimated. However, conclusive biochemical evidence on this is
missing, which will be imperative for the comprehensive design of potential enzyme
inhibitors.
Peptidoglycan breakdown and recycling
Most bacteria possess numerous enzymes capable of cleaving covalent bonds within the
peptidoglycan cell wall and/or its fragments, collectively known as peptidoglycan
hydrolases [44]. They play critical roles in the regulation of cell wall growth and turnover,
13
Peptidoglycan chain
TP
Donor/moenomycin
binding site
TG
Lipid II
binding site
Moenomycin
Figure 4. Lipid II polymerization by transglycosylases. Transglycosylases contain an extended
binding pocket to accommodate Lipid II and the growing chain. Linear glycan strands are formed that
are incorporated into the existing cell wall via transpeptidation (TP). The molecular structure of
moenomycin is depicted.
cell separation during division [45] and autolysis [46, 47]. Many bacteria have a large
number of hydrolases and a majority hereof is redundant in function, complicating the
attribution of a defined function for each individual hydrolase. The peptidoglycan
hydrolases can generally be classed into three main groups: glycosidases, (carboxy- and
endo)peptidases and N-acetylmuramyl-L-alanine amidases (Fig. 5). The glycosidases
cleave glycosidic bonds within the glycan strand, with lysozymes and lytic
transglycosidases specifically cleaving the β1-4 bond between MurNAc and GlcNAc in
different ways. The endo-N-acetyl-β-D-glucosaminidases more generally hydrolyze the
glycosidic bond between GlcNAc and a neighboring sugar in a variety of substrates
including peptidoglyan.
Carboxy- and endopeptidases cleave amide bonds between amino acids. Carboxypeptidases
specifically cleave off C-terminal amino acids. In E. coli, the D,D-carboxypeptidase PBP5
14
Figure 5. Peptidoglycan hydrolases: Glycosidases (1) cleave the glycosidic bond between MurNAc
and GlcNAc. N-acetylmuramyl-L-alanine amidases (2) hydrolyse the bond between MurNAc and Lalanine, releasing the peptide from the glycan chain. Carboxypeptidases (3) cleave peptide bonds to
remove C-terminal amino acids, most commonly the terminal D-alanine from the pentapeptide.
Endopeptidases (4) hydrolyze amide bonds within the peptides, resulting in cleavage of peptide
crosslinks.
cleaves off the terminal D-Ala from pentapeptides to form tetrapeptides. This changes the
crosslinking properties of the substrate as pentapeptides can act as both donors and
acceptors, whereas tetrapeptides only as acceptors. The activity of PBP5 seems to be
critical for the correct orientation of the septum and regular cell shape [48]. Endopeptidases
cleave within the peptide, most notably between bonds in peptide crossbridges. E. coli
PBP4 and PBP7 are endopeptidases.
N-acetylmuramyl-L-alanine amidases (amidases) cleave the amide bond between the
peptide and MurNAc thereby removing peptide side chains from the carbohydrate polymer
[10]. As with the other classes of hydrolases, multiple amidases are often present in a
bacterial species. E. coli has five: the homologous AmiA, AmiB and AmiC, which are
15
located in the periplasm, the cytoplasmic AmpD and the recently discovered AmiD, located
in the outer membrane [49]. AmpD and AmiD are also structurally similar and utilize
anhydroMurNAc-peptides as substrates (see next section) although AmiD was shown to
have a somewhat broader specificity.
Cells lacking multiple hydrolases show severe separation defects during cell division.
Judging from phenotypes of the various hydrolase mutants, the amidases, of which only
AmiC actually has been shown to localize to the septal ring during division [50], appear to
play the predominant part [51], although a recent report attributed only a minor role for the
amidases during cell separation [52]. E. coli mutants lacking all three amidases are unable
to divide and grow in filaments with thick septal peptidoglycan rings at division sites,
whereas these effects are diminished when any one of the amidases is still active, indicating
that their functions are redundant [45, 53].
Some Gram-negative bacteria, such as E. coli, reutilize all the breakdown products of their
peptidoglycan [54]. The exact function of this peptidoglycan recycling is still not clear. It is
not essential under growth conditions in the laboratory and none of the gene products
participating in peptidoglycan recycling seem to be required for viability, regular
morphology or growth rate. It could be a means to provide the cell with necessary nutrients
in the stationary phase or under starvation conditions. Moreover, recycling intermediates
cause the induction of ß-lactamase expression, which might be an indication that they play a
signaling role through which the condition of the cell wall and its metabolism can be
monitored [55, 56]. As peptidoglycan fragments are known to be potent stimulators of the
innate immune system, prevention of their release through recycling could also comprise an
escape mechanism from these immune responses [57].
Cleavage of the glycan strand between MurNAc and GlcNAc residues by lytic
transglycosylases concomitantly forms 1,6 anhydroMurNAc. GlcNAc-anhMurNAcpeptides are transported into the cytoplasm via AmpG, which is the main permease
involved in peptidoglycan recycling. In the cytoplasm, they are degraded by the Nacetylglucosaminidase (NagZ), anhydro-N-acetylmuramyl-L-Ala amidase (AmpD) and the
LD-carboxypeptidase (LdcA) forming the key recycling intermediate murein tripeptide (Lalanine-γ-D-glutamine-meso-diaminopimelic acid) (Fig. 6)[58].
Alternatively, peptide turnover products are to a minor extent produced in the periplasm
through amidase activity. Murein tripeptide obtained in the periplasm this way is recycled
16
by making use of the oligopeptide permease (Opp) pathway [58]. This is the main uptake
system for small peptides in E. coli which has a very broad specificity [59]. Initially,
murein tripeptide binds specifically to the periplasmic protein MppA (murein peptide
permease A) which transports the tripeptide to membrane-bound Opp components for
transport into the cytoplasm [60]. Here, a slight amount of murein tripeptide is degraded
further, however, the vast majority is directly ligated to UDP-MurNAc by murein peptide
ligase (Mpl). It thereby reenters the peptidoglycan biosynthesis route as UDP-MurNActripeptide that is the common substrate for MurF (Fig. 7) [61]. In vitro studies have shown
that Mpl is also able to utilize tetra- and pentapeptides as well as lysine containing peptides
[62]. The amino sugars GlcNAc and anhMurNAc freed in the cytoplasm are also recycled
for various purposes. Both are converted by different enzymes into GlcNAc-6-P, which can
serve as a precursor of UDP-GlcNAc in peptidoglycan and lipopolysaccharide biosynthesis,
and/or which can be metabolized by glycolysis.
Figure 6. Peptidoglycan recycling pathway. Breakdown products, by the action of peptidoglycan
hydrolases, are taken up into the cytoplasm via several permeases, where they are further processed.
Mpl attaches murein tripeptide (L-Ala-D-Glu-m-DAP) to UDP-MurNAc, yielding UDP-MurNActripeptide, which re-enters the peptidoglycan biosynthesis pathway as a substrate for MurF.
17
Use of labeled peptidoglycan derivatives in studying cell wall biosynthesis and its
inhibition
The cell wall synthesis machinery shows a high tolerance towards unnatural substrates.
MraY and MurG display broad substrate specificity since these proteins proved capable of
Lipid II synthesis utilizing (fluorescently) labeled substrates as well as prenyl carriers with
prenyl chain lengths varying from 2–25 isoprene units [63, 64]. Bacteria also tolerate
certain unnatural amino acids in their cell wall. For instance, mutants of E. coli lacking a
diaminopimelate epimerase were found to have a larger amount of L,L-diaminopimelic acid
(L,L-DAP) than meso-DAP incorporated into their peptidoglycan layer [65, 66]. Moreover,
in vivo experiments showed that DAP could be totally replaced by meso-lanthionine or Lallo-cystationine in peptidoglycan of E. coli [67]. As mentioned earlier, a wide variety of
amino acids can be found in the peptide stem of peptidoglycan throughout the bacterial
kingdom, which could explain the flexibility of the biosynthesis machinery towards
diversity in this part.
Many pioneering studies on the biosynthesis of peptidoglycan involved the use of
radiolabeled precursors [68-70], which were employed in studies focusing on the
elucidation of the mode of action of different antibiotics [71, 72] and the discovery of the
recycling pathway [58]. Beside their usefulness for analyses of bacterial cell wall
metabolism, radiolabeled peptidoglycan and its precursors are also of great interest for
investigations of host-bacteria interactions. For example, the injection of [14C]-DAP-labeled
peptidoglycan fragments in Drosophila allowed for the demonstration of an important role
of the amidase activity of a peptidoglycan-recognition protein in the response to bacterial
infection in the innate immune response in flies [73]. Although radiolabeled cell wall
precursors remain a valuable tool in cell wall studies, the high flexibility of the bacterial
cell wall synthesis machinery opened up the possibility to employ precursors modified by a
wide variety of labels. This drastically expanded the possibilities to study cell wall
synthesis and inhibitors thereof in more detail. One of the earliest studies using labeled
peptidoglycan derivatives employed UDP-MurNAc pentapeptide that was chemically
modified at its lysine residue with the spin-label Tempyo (2,2,5,5-tetramethyl-Noxylpyrroline-3-carbonyl) [74, 75]. A spin label provides information about its
microenvironment and mobility, and the labeled UDP-MurNAc-pentapeptide was used to
study the interactions with the antibiotics vancomycin and ristocetin during various stages
18
of in vitro peptidoglycan synthesis. Labeled cell wall precursor derivatives have also been
utilized as an effective tool to study surface display on living bacteria. UDP-MurNAcpentapeptide was modified with fluorescein or a ketone group. Bacteria were cultured in the
presence of the modified precursors, and the groups were displayed on the bacterial surface
via the cell wall biosynthesis machinery [76-78]. For incorporation of the precursors in
Gram-negative bacteria, additional EDTA treatment was required to permeabilize the outer
membrane. When Lactobacilli were treated with the ketone-modified precursor,
oligomannose was coupled with the ketone moiety on the bacterial surface through an
aminooxyl linker. The adhesion of these cells onto a lectin-containing surface increased
significantly compared to that of native bacteria. Since bacterial adhesion is an important
event in the infection of host cells and in the interaction between bacteria, controlling the
adhesion properties of the bacterial surface could provide benefits to the development of
novel bacterial drugs [79].
At a more detailed level, fluorescently labeled cell wall precursors have been extensively
exploited in order to study the reactions catalyzed by enzymes in the lipid-linked cycle of
bacterial peptidoglycan biosynthesis. As the quantum yield of certain fluorophores
increases significantly in hydrophobic surroundings, enzymatic reactions involving changes
in the hydrophobic environment of specific substrates are ideally suited for fluorescence
assays. The molecular basis for the inhibition of MraY from E. coli by two classes of
antibiotics, the mureidomycins and the liposidomycins, was assayed with a fluorescencebased enzyme assay [80, 81]. UDP-MurNAc-pentapeptide, dansylated at its DAP, was used
as a substrate. This allowed for the kinetic characterization of mureidomycin A,
tunicamycin and liposidomycin B. This method was subsequently expanded to a highthroughput assay to screen for inhibitors of MraY [82].
A high-throughput approach was also reported to identify glycosyltransferase inhibitors.
This involved the displacement of a UDP-GlcNAc derivative, fluorescently modified at its
N-acetyl group, from the glycosyl donor binding site, and this screen was successfully
applied to E. coli MurG [83, 84]. Furthermore, an assay based on fluorescence resonance
energy transfer (FRET) was developed for MurG [85]. In this way, the kinetic parameters
for the enzymatic processing by MurG were determined with UDP-N-acetylglucosamine,
fluorescently labeled at the C-6 position with indole-3-acetic acid and Lipid I labeled with
dansyl at the DAP residue. Aside from being a valuable tool for interaction studies and
19
inhibitor screenings, chemically modified precursors themselves could be promising
candidates for a new type of antibiotic. Fluorinated carbohydrate derivatives are extensively
applied in medical sciences as substrate mimics for the inhibition of enzymatic processes,
because most enzymes are not able to differentiate between the fluorinated and original
compounds. Lipid I analogues, fluorinated at the C-4 position of the MurNAc residue, were
shown to be potent inhibitors of MurG [86], and when Gram-positive bacteria were
incubated in the presence of UDP-MurNAc pentapeptide fluorinated at the position C-4,
significant inhibition of growth was observed [87].
The investigation of enzymes utilizing Lipid I and Lipid II has long been hindered by the
difficulty of acquiring these substrates in useful quantities from natural sources. In an effort
to overcome this problem, several groups have now reported the chemical synthesis of
Lipid II [64, 88, 89]. Alternatively, a semi-synthetic approach was developed to synthesize
large quantities of Lipid II and variants [63]. This latter approach involves supplying
bacterial membrane preparations rich in MraY and MurG with polyisoprenylphosphates,
UDP-MurNAc-pentapeptide and UDP-GlcNAc. Fluorescent Lipid II was produced with
UDP-MurNAc-pentapeptide modified at its lysine residue with a fluorescent label. Assays
were developed to study the polymerization of a fluorescently labeled Lipid II derivative by
E. coli PBP1b [90] and the glycosyltransferase domains of Thermotoga maritima PBP1a
[91] and Streptococcus pneumoniae PBP2a [92].
Another important application of labeling the cell wall is in the research on peptidoglycan
metabolism throughout the bacterial cell cycle. Peptidoglycan segregation was studied
using a labeling approach unlike the aforementioned ones, i.e. the ability of E. coli to
incorporate D-cysteine through an unclear mechanism of periplasmic amino acid exchange
into its peptidoglycan [93]. Following the biotinylation of the cysteine thiol groups, the
distribution of modified peptidoglycan in purified peptidoglycan sacculi could be traced
and visualized by immunodetection with fluorescence and electron microscopy techniques.
Using pulse-chase to differentiate between old (labeled) and new peptidoglycan, it was
shown that at the initiation of cell division, a narrow zone at the future division site was
devoid of labels, suggesting that all peptidoglycan here is newly synthesized. Moreover, the
amount of labeled peptidoglycan in polar caps was shown to be stable for many
generations, indicating that it is metabolically inert. This method was further applied to
20
reveal mechanisms of bacterial morphogenesis, such as the role of PBP5 in the correct
placement of inert peptidoglycan [94-96].
Divisome
Cell growth and division are closely coordinated with the synthesis and hydrolysis of
peptidoglycan. In rod-shaped bacteria this involves the elongation of the cylindrical part of
the sacculus and the formation of two new cell poles. During cell elongation, peptidoglycan
precursors are incorporated into the existing sacculus at a limited number of homogenously
distributed and highly mobile sites covering the cylindrical cell surface. This is performed
by a multi-enzyme complex, which uses the helical MreB cytoskeleton as a tethering
device. After the cell has grown to sufficient length and the DNA has been replicated, cell
division (binary fission) commences with the formation of a septum at mid-cell.
Constriction of the cell wall and cell membranes results in the formation of two daughter
cells [97-99]. Bacterial cell division requires the concerted action of a large group of
enzymes, collectively known as the divisome (Fig. 7). The exact function of most of these
proteins have yet to be determined [100]. Cytokinesis is initiated by the polymerization of
the tubulin homologue FtsZ into a ring-like structure (Z-ring) at mid-cell. This event
orchestrates the step-wise assembly of the other cell division proteins. Upon Z-ring
formation, FtsA, ZipA, ZapA and probably FtsE/X are recruited while a switch in the mode
of peptidoglycan synthesis occurs from dispersed in the lateral wall to localized at the
division site [101-104]. After a short time lag, the late division proteins FtsK, FtsQ, FtsL,
FtsB, FtsW, FtsI (PBP3), FtsN and AmiC are recruited to the division site [104, 105] and
septal peptidoglycan is ultimately shaped into the two new cell poles [53, 103, 104].
Following division, the peptidoglycan at the newly formed poles is believed to be
metabolically inert [93]. Elongation and division have for long been viewed as strictly
separated events, however recent evidence has revealed that FtsZ is also involved in
peptidoglycan synthesis during elongation [94, 106].
Cell wall proteomics
The close coordination between cell wall metabolism and progression of the bacterial cell
cycle requires a vast network of proteins. It involves dynamic and complex interactions of
cell wall-interacting proteins within often enormous multi-enzyme complexes. The
interactions and functions of many of the proteins involved are poorly understood.
21
AmiC
FtsN
ZipA
FtsX FtsK
FtsW MraY
FtsE
MurG
FtsA
FtsK
ZapA
FtsZ
Figure 7. Schematic representation of the E. coli division protein complex (divisome). The
compartmentalization and the supposed interactions of the proteins that localize at mid-cell during
cell division are depicted.
Furthermore, numerous players have yet to be identified. It is therefore essential to fully
elucidate the cell wall proteome and get a better understanding of enzyme structure,
function and dynamics. This is also likely to expand the range of potential drug targets.
Proteomics is the comprehensive study of proteins, which aims to provide a complete
description of the entire protein spectrum underlying cell physiology. Proteomic studies are
generally classified as gel-based, gel-independent or predictive. The former two categories
are defined as being based on separation of protein mixtures in gels and in a gel-free
environment,
respectively,
whereas
predictive
proteomics
entails
studies
using
computational methods [107, 108]. Gel-based approaches are by far the most widely
implemented. Separation of proteins is usually achieved on polyacrylamide gels in either
one (SDS-PAGE) or two dimensions. Two-dimensional gel electrophoresis (2DE) is the
22
method of choice for the analysis of complex protein mixtures. It couples separation
according to isoelectric points, termed isoelectric focusing (IEF), in the first dimension with
size-based SDS-PAGE in the second. The development of improved methods for protein
detection has facilitated the advancement of gel-based proteomics to a great extent.
Coomassie brilliant blue and silver staining have been the typical detection methods, while
the superior sensitivity of radiolabeling is offset by obvious hazards and high costs. The use
of fluorescent dyes has gained popularity due to their higher sensitivity and wider dynamic
range compared to the colorimetric procedures. An example of fluorescence-based
detection is differential in gel electrophoresis (DIGE), where multiple protein samples are
each labeled with a different dye and simultaneously separated. Successive excitation of
each dye allows for the analysis of differentially expressed proteins in a single run [109].
Following separation and detection, typically, protein-containing gel pieces are excised and
the proteins digested with a protease, most often trypsin. The resulting peptides are
subsequently analyzed by tandem mass spectrometry and matched to theoretical peptide
masses in protein sequence databases. The major drawback of the gel-based technologies
lies in their inability to deal with certain classes of proteins, in particular very hydrophobic
ones, like membrane and cytoskeletal proteins, and those with highly acidic or basic
isoelectric points. Due to these limitations non-gel separation approaches have been
developed. The most significant methodology couples multidimensional chromatographic
separation of peptides from a complex digest mixture to MS identification.
Multidimensional protein identification technology (MudPIT) is probably the most popular
variant, using strong cation-exchange in the first dimension followed by reversed-phase
chromatography in the second. The cell wall proteomes of two Listeria strains were
analyzed using MudPIT, after exploiting the insolubility of peptidoglycan in boiling SDS
solutions for prefractionation purposes [110]. The continuous advances in performance of
mass analyzers, such as the recently developed hybrid Orbitrap instruments, allow the
identification of vast numbers of proteins with high mass accuracy and resolution as was
shown in a large-scale proteomics analysis of Mycobacterium tuberculosis [111]. Next to
optimization of separation and detection techniques, the complexity of biological analytes
also demands continuous development of enrichment strategies. Selective labeling of
proteins via substrate analogues has become a powerful tool for the profiling of enzyme
activities [112], enzyme-substrate interactions [113] and post-translational modifications
[114]. Key advances in this field have been facilitated by the emergence of the bio-
23
orthogonal click chemistry reaction, the 1,3-dipolar cycloaddition between an azide and an
alkyne forming a stable triazole product, as a means to attach reporter tags subsequent to
protein labeling [112, 115]. This approach has been particularly beneficial by allowing high
access to biologically more relevant systems, such as biological membranes and living
cells. Small alkyne-derivatized β-lactams were synthesized and exploited in vivo to capture
enzymes involved in cell wall synthesis, antibiotic resistance and virulence [116]. Due to its
notorious multi-resistant strains, Staphylococcus aureus has been the subject of numerous
cell wall-related proteomic studies. Changes in protein abundance were profiled in strains
with different levels of vancomycin resistance [117], whereas cell wall-associated proteins
were mapped and their potential as vaccine candidates was addressed [118, 119].
Immunogenic cell wall proteins were also identified in Clostridium difficile [120] and
Streptococcus suis [121].
Scope of the thesis
This thesis describes the development and application of novel techniques to gain more
insight into peptidoglycan metabolism and the enzymes involved herein with an emphasis
on the role of Lipid II in the targeting of transglycosylases. The high tolerance of the
bacterial cell wall synthesis machinery is fully exploited by the use of a variety of labeled
peptidoglycan derivatives. In Chapter 2, a fluorescent cell wall labeling approach is set up,
which is used to study peptidoglycan metabolism in vivo. It is shown here that externally
supplied NBD-labeled murein tripeptide is taken up by E. coli cells and metabolically
incorporated into the cell wall via the peptidoglycan recycling pathway. By analyzing the
cell wall labeling patterns in wild-type cells and several division and amidase mutants,
FtsZ-dependent
hydrolase
activity
during
preseptal
elongation
was
discovered.
Additionally, we could visualize the major peptidoglycan hydrolase activity of AmiC
during septation. In Chapter 3, the above developed method of labeling the cell wall of E.
coli with reporter groups in vivo is adapted into a proteomics format. Using a tripeptide
derivative containing a photoactivatable crosslinker and an alkyne moiety, proteins
interacting with the cell wall and/or its precursors were crosslinked. Using the alkyne as
bait, azide derivatized tags or beads could be covalently attached to the crosslink products
via click chemistry, enabling their selective detection and purification. A number of
interesting proteins were identified. In Chapters 4 & 5, the focus is on the
transglycosylation process, which for long was a black box in peptidoglycan biosynthesis
24
and a potentially interesting target for novel antibiotics. In Chapter 4, in vitro photocrosslinking in combination with mass spectrometry techniques provided information on
the substrate (Lipid II) binding site in the transglycosylase penicillin-binding protein 1b
(PBP1b) of E. coli. By reacting a photoactivatable analogue of Lipid II with the enzyme in
the presence of moenomycin, it is shown that the substrate is covalently captured in one
specific region, possibly encompassing the binding site of Lipid II. In Chapter 5, the
mechanism of action of the potential transglycosylase inhibitor compound 5b is
investigated. 5b is shown to disturb the functional integrity of negatively charged
membranes. Moreover, nisin-Lipid II pore formation in model membranes was inhibited by
5b, which was reduced by an excess presence of undecaprenylpyrophosphate. This points to
a specific interaction of 5b with Lipid II. In Chapter 6 the obtained results are recapitulated
and discussed.
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
Razin, S., D. Yogev, and Y. Naot, Molecular biology and pathogenicity of
mycoplasmas. Microbiol Mol Biol Rev, 1998. 62(4): p. 1094-156.
Gan, L., S. Chen, and G.J. Jensen, Molecular organization of Gram-negative
peptidoglycan. Proc Natl Acad Sci U S A, 2008. 105(48): p. 18953-7.
Braun, V. and K. Rehn, Chemical characterization, spatial distribution and
function of a lipoprotein (murein-lipoprotein) of the E. coli cell wall. The specific
effect of trypsin on the membrane structure. Eur J Biochem, 1969. 10(3): p. 42638.
Leduc, M., et al., Interactions of Escherichia coli membrane lipoproteins with the
murein sacculus. J Bacteriol, 1992. 174(24): p. 7982-8.
Clavel, T., et al., TolB protein of Escherichia coli K-12 interacts with the outer
membrane peptidoglycan-associated proteins Pal, Lpp and OmpA. Mol Microbiol,
1998. 29(1): p. 359-67.
Schleifer, K.H. and O. Kandler, Peptidoglycan types of bacterial cell walls and
their taxonomic implications. Bacteriol Rev, 1972. 36(4): p. 407-77.
Ghuysen, J.M., Use of bacteriolytic enzymes in determination of wall structure
and their role in cell metabolism. Bacteriol Rev, 1968. 32(4 Pt 2): p. 425-64.
Vollmer, W., Structural variation in the glycan strands of bacterial peptidoglycan.
FEMS Microbiol Rev, 2008. 32(2): p. 287-306.
Boniface, A., et al., The MurE synthetase from Thermotoga maritima is endowed
with an unusual D-lysine adding activity. J Biol Chem, 2006. 281(23): p. 15680-6.
Holtje, J.V., Growth of the stress-bearing and shape-maintaining murein sacculus
of Escherichia coli. Microbiol Mol Biol Rev, 1998. 62(1): p. 181-203.
Brown, E.D., et al., MurA (MurZ), the enzyme that catalyzes the first committed
step in peptidoglycan biosynthesis, is essential in Escherichia coli. J Bacteriol,
1995. 177(14): p. 4194-7.
Mengin-Lecreulx, D., B. Flouret, and J. van Heijenoort, Cytoplasmic steps of
peptidoglycan synthesis in Escherichia coli. J Bacteriol, 1982. 151(3): p. 1109-17.
25
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
Barreteau, H., et al., Cytoplasmic steps of peptidoglycan biosynthesis. FEMS
Microbiol Rev, 2008. 32(2): p. 168-207.
Mizuno, Y., M. Yaegashi, and E. Ito, Purification and properties of uridine
diphosphate N-acetylmuramate: L-alanine ligase. J Biochem, 1973. 74(3): p. 52538.
Pucci, M.J., L.F. Discotto, and T.J. Dougherty, Cloning and identification of the
Escherichia coli murB DNA sequence, which encodes UDP-Nacetylenolpyruvoylglucosamine reductase. J Bacteriol, 1992. 174(5): p. 1690-3.
Bouhss, A., et al., The biosynthesis of peptidoglycan lipid-linked intermediates.
FEMS Microbiol Rev, 2008. 32(2): p. 208-33.
van Heijenoort, J., Lipid intermediates in the biosynthesis of bacterial
peptidoglycan. Microbiol Mol Biol Rev, 2007. 71(4): p. 620-35.
Mohammadi, T., et al., Identification of FtsW as a transporter of lipid-linked cell
wall precursors across the bacterial cytoplasmic membrane. submitted.
Ikeda, M., et al., Structural similarity among Escherichia coli FtsW and RodA
proteins and Bacillus subtilis SpoVE protein, which function in cell division, cell
elongation, and spore formation, respectively. J Bacteriol, 1989. 171(11): p. 63758.
El Ghachi, M., et al., The bacA gene of Escherichia coli encodes an undecaprenyl
pyrophosphate phosphatase activity. J Biol Chem, 2004. 279(29): p. 30106-13.
El Ghachi, M., et al., Identification of multiple genes encoding membrane proteins
with undecaprenyl pyrophosphate phosphatase (UppP) activity in Escherichia
coli. J Biol Chem, 2005. 280(19): p. 18689-95.
Touze, T., D. Blanot, and D. Mengin-Lecreulx, Substrate specificity and
membrane topology of Escherichia coli PgpB, an undecaprenyl pyrophosphate
phosphatase. J Biol Chem, 2008. 283(24): p. 16573-83.
Valvano, M.A., Undecaprenyl phosphate recycling comes out of age. Mol
Microbiol, 2008. 67(2): p. 232-5.
Touze, T., et al., Periplasmic phosphorylation of lipid A is linked to the synthesis
of undecaprenyl phosphate. Mol Microbiol, 2008. 67(2): p. 264-77.
Sauvage, E., et al., The penicillin-binding proteins: structure and role in
peptidoglycan biosynthesis. FEMS Microbiol Rev, 2008. 32(2): p. 234-58.
Ishino, F., et al., Dual enzyme activities of cell wall peptidoglycan synthesis,
peptidoglycan transglycosylase and penicillin-sensitive transpeptidase, in purified
preparations of Escherichia coli penicillin-binding protein 1A. Biochem Biophys
Res Commun, 1980. 97(1): p. 287-93.
Nakagawa, J., et al., Functional biosynthesis of cell wall peptidoglycan by
polymorphic bifunctional polypeptides. Penicillin-binding protein 1Bs of
Escherichia coli with activities of transglycosylase and transpeptidase. J Biol
Chem, 1984. 259(22): p. 13937-46.
Henrichfreise, B., et al., Functional conservation of the lipid II biosynthesis
pathway in the cell wall-less bacteria Chlamydia and Wolbachia: why is lipid II
needed? Mol Microbiol, 2009. 73(5): p. 913-23.
Arthur, M., et al., Evidence for in vivo incorporation of D-lactate into
peptidoglycan precursors of vancomycin-resistant enterococci. Antimicrob Agents
Chemother, 1992. 36(4): p. 867-9.
Guskey, M.T. and B.T. Tsuji, A comparative review of the lipoglycopeptides:
oritavancin, dalbavancin, and telavancin. Pharmacotherapy. 30(1): p. 80-94.
26
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
Cudic, P., et al., Complexation of peptidoglycan intermediates by the
lipoglycodepsipeptide antibiotic ramoplanin: minimal structural requirements for
intermolecular complexation and fibril formation. Proc Natl Acad Sci U S A,
2002. 99(11): p. 7384-9.
Hasper, H.E., et al., An alternative bactericidal mechanism of action for lantibiotic
peptides that target lipid II. Science, 2006. 313(5793): p. 1636-7.
Breukink, E. and B. de Kruijff, Lipid II as a target for antibiotics. Nat Rev Drug
Discov, 2006. 5(4): p. 321-32.
Bonev, B.B., et al., Targeting extracellular pyrophosphates underpins the high
selectivity of nisin. FASEB J, 2004. 18(15): p. 1862-9.
Hasper, H.E., B. de Kruijff, and E. Breukink, Assembly and stability of nisin-lipid
II pores. Biochemistry, 2004. 43(36): p. 11567-75.
Smith, L., et al., Elucidation of the antimicrobial mechanism of mutacin 1140.
Biochemistry, 2008. 47(10): p. 3308-14.
Derouaux, A., et al., The monofunctional glycosyltransferase of Escherichia coli
localizes to the cell division site and interacts with penicillin-binding protein 3,
FtsW, and FtsN. J Bacteriol, 2008. 190(5): p. 1831-4.
Breukink, E., et al., Use of the cell wall precursor lipid II by a pore-forming
peptide antibiotic. Science, 1999. 286(5448): p. 2361-4.
Halliday, J., et al., Targeting the forgotten transglycosylases. Biochem Pharmacol,
2006. 71(7): p. 957-67.
Yuan, Y., et al., Crystal structure of a peptidoglycan glycosyltransferase suggests
a model for processive glycan chain synthesis. Proc Natl Acad Sci U S A, 2007.
104(13): p. 5348-53.
Lovering, A.L., et al., Structural insight into the transglycosylation step of
bacterial cell-wall biosynthesis. Science, 2007. 315(5817): p. 1402-5.
Sung, M.T., et al., Crystal structure of the membrane-bound bifunctional
transglycosylase PBP1b from Escherichia coli. Proc Natl Acad Sci U S A, 2009.
106(22): p. 8824-9.
Yuan, Y., et al., Structural analysis of the contacts anchoring moenomycin to
peptidoglycan glycosyltransferases and implications for antibiotic design. ACS
Chem Biol, 2008. 3(7): p. 429-36.
Vollmer, W., et al., Bacterial peptidoglycan (murein) hydrolases. FEMS
Microbiol Rev, 2008. 32(2): p. 259-86.
Heidrich, C., et al., Effects of multiple deletions of murein hydrolases on viability,
septum cleavage, and sensitivity to large toxic molecules in Escherichia coli. J
Bacteriol, 2002. 184(22): p. 6093-9.
Antignac, A., K. Sieradzki, and A. Tomasz, Perturbation of cell wall synthesis
suppresses autolysis in Staphylococcus aureus: evidence for coregulation of cell
wall synthetic and hydrolytic enzymes. J Bacteriol, 2007. 189(21): p. 7573-80.
Yamada, S., et al., An autolysin ring associated with cell separation of
Staphylococcus aureus. J Bacteriol, 1996. 178(6): p. 1565-71.
Popham, D.L. and K.D. Young, Role of penicillin-binding proteins in bacterial
cell morphogenesis. Curr Opin Microbiol, 2003. 6(6): p. 594-9.
Uehara, T. and J.T. Park, An anhydro-N-acetylmuramyl-L-alanine amidase with
broad specificity tethered to the outer membrane of Escherichia coli. J Bacteriol,
2007. 189(15): p. 5634-41.
27
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
Bernhardt, T.G. and P.A. de Boer, The Escherichia coli amidase AmiC is a
periplasmic septal ring component exported via the twin-arginine transport
pathway. Mol Microbiol, 2003. 48(5): p. 1171-82.
Heidrich, C., et al., Involvement of N-acetylmuramyl-L-alanine amidases in cell
separation and antibiotic-induced autolysis of Escherichia coli. Mol Microbiol,
2001. 41(1): p. 167-78.
Uehara, T. and J.T. Park, Growth of Escherichia coli: significance of
peptidoglycan degradation during elongation and septation. J Bacteriol, 2008.
190(11): p. 3914-22.
Priyadarshini, R., de Pedro, M.A. and Young, K.D. , Role of peptidoglycan
amidases in the development and morphology of the division septum in
Escherichia coli. . J Bacteriol, 2007(in press).
Park, J.T. and T. Uehara, How bacteria consume their own exoskeletons (turnover
and recycling of cell wall peptidoglycan). Microbiol Mol Biol Rev, 2008. 72(2): p.
211-27, table of contents.
Park, J.T., Why does Escherichia coli recycle its cell wall peptides? Mol
Microbiol, 1995. 17(3): p. 421-6.
Jacobs, C., et al., Bacterial cell wall recycling provides cytosolic muropeptides as
effectors for beta-lactamase induction. Embo J, 1994. 13(19): p. 4684-94.
Mengin-Lecreulx, D. and B. Lemaitre, Structure and metabolism of peptidoglycan
and molecular requirements allowing its detection by the Drosophila innate
immune system. J Endotoxin Res, 2005. 11(2): p. 105-11.
Goodell, E.W. and C.F. Higgins, Uptake of cell wall peptides by Salmonella
typhimurium and Escherichia coli. J Bacteriol, 1987. 169(8): p. 3861-5.
Higgins, C.F. and M.M. Gibson, Peptide transport in bacteria. Methods Enzymol,
1986. 125: p. 365-77.
Park, J.T., et al., MppA, a periplasmic binding protein essential for import of the
bacterial cell wall peptide L-alanyl-gamma-D-glutamyl-meso-diaminopimelate. J
Bacteriol, 1998. 180(5): p. 1215-23.
Mengin-Lecreulx, D., J. van Heijenoort, and J.T. Park, Identification of the mpl
gene encoding UDP-N-acetylmuramate: L-alanyl-gamma-D-glutamyl-mesodiaminopimelate ligase in Escherichia coli and its role in recycling of cell wall
peptidoglycan. J Bacteriol, 1996. 178(18): p. 5347-52.
Herve, M., et al., Biochemical characterization and physiological properties of
Escherichia coli UDP-N-acetylmuramate:L-alanyl-gamma-D-glutamyl-mesodiaminopimelate ligase. J Bacteriol, 2007. 189(11): p. 3987-95.
Breukink, E., et al., Lipid II is an intrinsic component of the pore induced by nisin
in bacterial membranes. J Biol Chem, 2003. 278(22): p. 19898-903.
Ye, X.Y., et al., Better substrates for bacterial transglycosylases. J Am Chem Soc,
2001. 123(13): p. 3155-6.
Mengin-Lecreulx, D., et al., Incorporation of LL-diaminopimelic acid into
peptidoglycan of Escherichia coli mutants lacking diaminopimelate epimerase
encoded by dapF. J Bacteriol, 1988. 170(5): p. 2031-9.
Richaud, C., et al., Molecular cloning, characterization, and chromosomal
localization of dapF, the Escherichia coli gene for diaminopimelate epimerase. J
Bacteriol, 1987. 169(4): p. 1454-9.
Richaud, C., et al., Directed evolution of biosynthetic pathways. Recruitment of
cysteine thioethers for constructing the cell wall of Escherichia coli. J Biol Chem,
1993. 268(36): p. 26827-35.
28
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
Anderson, J.S., et al., Lipid-Phosphoacetylmuramyl-Pentapeptide and LipidPhosphodisaccharide-Pentapeptide:
Presumed
Membrane
Transport
Intermediates in Cell Wall Synthesis. Proc Natl Acad Sci U S A, 1965. 53: p. 8819.
Hammes, W.P. and F.C. Neuhaus, On the specificity of phospho-N-acetylmuramylpentapeptide translocase. The peptide subunit of uridine diphosphate-Nactylmuramyl-pentapeptide. J Biol Chem, 1974. 249(10): p. 3140-50.
Higashi, Y., J.L. Strominger, and C.C. Sweeley, Structure of a lipid intermediate
in cell wall peptidoglycan synthesis: a derivative of a C55 isoprenoid alcohol.
Proc Natl Acad Sci U S A, 1967. 57(6): p. 1878-84.
Siewert, G. and J.L. Strominger, Bacitracin: An Inhibitor of the
Dephosphorylation of Lipid Pyrophosphate, an Intermediate in the Biosynthesis of
the Peptidoglycan of Bacterial Cell Walls. Proc Natl Acad Sci U S A, 1967. 57(3):
p. 767-773.
Ward, J.B., Tunicamycin inhibition of bacterial wall polymer synthesis. FEBS
Lett, 1977. 78(1): p. 151-4.
Zaidman-Remy, A., et al., The Drosophila amidase PGRP-LB modulates the
immune response to bacterial infection. Immunity, 2006. 24(4): p. 463-73.
Johnston, L.S. and F.C. Neuhaus, Initial membrane reaction in the biosynthesis of
peptidoglycan. Spin-labeled intermediates as receptors for vancomycin and
ristocetin. Biochemistry, 1975. 14(12): p. 2754-60.
Johnston, L.S. and F.C. Neuhaus, Biosynthesis of spin-labeled peptidoglycan:
spin-spin interactions. Biochemistry, 1977. 16(7): p. 1251-7.
Sadamoto, R., et al., Cell wall engineering of living bacteria through biosynthesis.
Methods Enzymol, 2003. 362: p. 273-86.
Sadamoto, R., et al., Cell-wall engineering of living bacteria. J Am Chem Soc,
2002. 124(31): p. 9018-9.
Sadamoto, R., et al., Control of bacteria adhesion by cell-wall engineering. J Am
Chem Soc, 2004. 126(12): p. 3755-61.
Sharon, N. and I. Ofek, Safe as mother's milk: carbohydrates as future antiadhesion drugs for bacterial diseases. Glycoconj J, 2000. 17(7-9): p. 659-64.
Brandish, P.E., et al., Slow binding inhibition of phospho-N-acetylmuramylpentapeptide-translocase (Escherichia coli) by mureidomycin A. J Biol Chem,
1996. 271(13): p. 7609-14.
Brandish, P.E., et al., Modes of action of tunicamycin, liposidomycin B, and
mureidomycin A: inhibition of phospho-N-acetylmuramyl-pentapeptide
translocase from Escherichia coli. Antimicrob Agents Chemother, 1996. 40(7): p.
1640-4.
Stachyra, T., et al., Fluorescence detection-based functional assay for highthroughput screening for MraY. Antimicrob Agents Chemother, 2004. 48(3): p.
897-902.
Helm, J.S., et al., Identification of active-site inhibitors of MurG using a
generalizable, high-throughput glycosyltransferase screen. J Am Chem Soc, 2003.
125(37): p. 11168-9.
Hu, Y., et al., Identification of selective inhibitors for the glycosyltransferase
MurG via high-throughput screening. Chem Biol, 2004. 11(5): p. 703-11.
Li, J.J. and T.D. Bugg, A fluorescent analogue of UDP-N-acetylglucosamine:
application for FRET assay of peptidoglycan translocase II (MurG). Chem
Commun (Camb), 2004(2): p. 182-3.
29
86.
87.
88.
89.
90.
91.
92.
93.
94.
95.
96.
97.
98.
99.
100.
101.
102.
103.
104.
Chen, L., et al., Intrinsic lipid preferences and kinetic mechanism of Escherichia
coli MurG. Biochemistry, 2002. 41(21): p. 6824-33.
Ueda, T., et al., Synthesis of 4-fluorinated UDP-MurNAc pentapeptide as an
inhibitor of bacterial growth. Org Lett, 2004. 6(11): p. 1753-6.
Schwartz, B., J.A. Markwalder, and Y. Wang, Lipid II: total synthesis of the
bacterial cell wall precursor and utilization as a substrate for glycosyltransfer and
transpeptidation by penicillin binding protein (PBP) 1b of Escherichia coli. J Am
Chem Soc, 2001. 123(47): p. 11638-43.
VanNieuwenhze, M.S., et al., The first total synthesis of lipid II: the final
monomeric intermediate in bacterial cell wall biosynthesis. J Am Chem Soc,
2002. 124(14): p. 3656-60.
Schwartz, B., et al., A kinetic characterization of the glycosyltransferase activity of
Eschericia coli PBP1b and development of a continuous fluorescence assay.
Biochemistry, 2002. 41(41): p. 12552-61.
Offant, J., et al., Functional characterization of the glycosyltransferase domain of
penicillin-binding protein 1a from Thermotoga maritima. Biochim Biophys Acta,
2006. 1764(6): p. 1036-42.
Di Guilmi, A.M., et al., The glycosyltransferase domain of penicillin-binding
protein 2a from Streptococcus pneumoniae catalyzes the polymerization of murein
glycan chains. J Bacteriol, 2003. 185(15): p. 4418-23.
de Pedro, M.A., et al., Murein segregation in Escherichia coli. J Bacteriol, 1997.
179(9): p. 2823-34.
Aaron, M., et al., The tubulin homologue FtsZ contributes to cell elongation by
guiding cell wall precursor synthesis in Caulobacter crescentus. Mol Microbiol,
2007. 64(4): p. 938-52.
de Pedro, M.A., et al., Constitutive septal murein synthesis in Escherichia coli
with impaired activity of the morphogenetic proteins RodA and penicillin-binding
protein 2. J Bacteriol, 2001. 183(14): p. 4115-26.
de Pedro, M.A., et al., Branching of Escherichia coli cells arises from multiple
sites of inert peptidoglycan. J Bacteriol, 2003. 185(4): p. 1147-52.
Canepari, P., et al., Cell elongation and septation are two mutually exclusive
processes in Escherichia coli. Arch Microbiol, 1997. 168(2): p. 152-9.
Nanninga, N., Morphogenesis of Escherichia coli. Microbiol Mol Biol Rev, 1998.
62(1): p. 110-29.
den Blaauwen, T., de Pedro, M. A., Nguyen-Disteche, M. and Ayala, J. A.,
Morphogenesis of the rodshaped sacculus. FEMS Microbiol Lett, 2008. in press.
Errington, J., R.A. Daniel, and D.J. Scheffers, Cytokinesis in bacteria. Microbiol
Mol Biol Rev, 2003. 67(1): p. 52-65, table of contents.
Rueda, S., M. Vicente, and J. Mingorance, Concentration and assembly of the
division ring proteins FtsZ, FtsA, and ZipA during the Escherichia coli cell cycle.
J Bacteriol, 2003. 185(11): p. 3344-51.
Corbin, B.D., et al., Interaction between cell division proteins FtsE and FtsZ. J
Bacteriol, 2007. 189(8): p. 3026-35.
Den Blaauwen, T., et al., Timing of FtsZ assembly in Escherichia coli. J Bacteriol,
1999. 181(17): p. 5167-75.
Aarsman, M.E., et al., Maturation of the Escherichia coli divisome occurs in two
steps. Mol Microbiol, 2005. 55(6): p. 1631-45.
30
105.
106.
107.
108.
109.
110.
111.
112.
113.
114.
115.
116.
117.
118.
119.
120.
121.
Wang, X., C. Possoz, and D.J. Sherratt, Dancing around the divisome: asymmetric
chromosome segregation in Escherichia coli. Genes Dev, 2005. 19(19): p. 236777.
Varma, A., M.A. de Pedro, and K.D. Young, FtsZ directs a second mode of
peptidoglycan synthesis in Escherichia coli. J Bacteriol, 2007. 189(15): p. 5692704.
VanBogelen, R.A., et al., Escherichia coli proteome analysis using the geneprotein database. Electrophoresis, 1997. 18(8): p. 1243-51.
Blattner, F.R., et al., The complete genome sequence of Escherichia coli K-12.
Science, 1997. 277(5331): p. 1453-62.
Unlu, M., M.E. Morgan, and J.S. Minden, Difference gel electrophoresis: a single
gel method for detecting changes in protein extracts. Electrophoresis, 1997.
18(11): p. 2071-7.
Calvo, E., et al., Analysis of the Listeria cell wall proteome by two-dimensional
nanoliquid chromatography coupled to mass spectrometry. Proteomics, 2005.
5(2): p. 433-43.
de Souza, G.A., et al., High accuracy mass spectrometry analysis as a tool to
verify and improve gene annotation using Mycobacterium tuberculosis as an
example. BMC Genomics, 2008. 9: p. 316.
Speers, A.E., G.C. Adam, and B.F. Cravatt, Activity-based protein profiling in vivo
using a copper(i)-catalyzed azide-alkyne [3 + 2] cycloaddition. J Am Chem Soc,
2003. 125(16): p. 4686-7.
Lin, M.Z. and L. Wang, Selective labeling of proteins with chemical probes in
living cells. Physiology (Bethesda), 2008. 23: p. 131-41.
Chang, P.V., et al., Metabolic labeling of sialic acids in living animals with
alkynyl sugars. Angew Chem Int Ed Engl, 2009. 48(22): p. 4030-3.
Best, M.D., Click chemistry and bioorthogonal reactions: unprecedented
selectivity in the labeling of biological molecules. Biochemistry, 2009. 48(28): p.
6571-84.
Staub, I. and S.A. Sieber, Beta-lactam probes as selective chemical-proteomic
tools for the identification and functional characterization of resistance associated
enzymes in MRSA. J Am Chem Soc, 2009. 131(17): p. 6271-6.
Pieper, R., et al., Comparative proteomic analysis of Staphylococcus aureus
strains with differences in resistance to the cell wall-targeting antibiotic
vancomycin. Proteomics, 2006. 6(15): p. 4246-58.
Glowalla, E., et al., Proteomics-based identification of anchorless cell wall
proteins as vaccine candidates against Staphylococcus aureus. Infect Immun,
2009. 77(7): p. 2719-29.
Taverna, F., et al., Characterization of cell wall associated proteins of a
Staphylococcus aureus isolated from bovine mastitis case by a proteomic
approach. Vet Microbiol, 2007. 119(2-4): p. 240-7.
Wright, A., et al., Immunoreactive cell wall proteins of Clostridium difficile
identified by human sera. J Med Microbiol, 2008. 57(Pt 6): p. 750-6.
Zhang, A., et al., Identification of immunogenic cell wall-associated proteins of
Streptococcus suis serotype 2. Proteomics, 2008. 8(17): p. 3506-15.
31
32
Chapter 2
A novel in vivo cell wall labeling
approach sheds new light on
peptidoglycan synthesis in
Escherichia coli.
Olrichs, N.K., Aarsman, M. E. G., Arnusch, C. J., Vollmer, W., de Kruijff, B., Breukink,
E., den Blaauwen, T. submitted.
Abstract
Peptidoglycan synthesis and turnover in relation to cell growth and division has been
studied using a new labeling method. This method involves the incorporation of
fluorescently labeled peptidoglycan precursors into the cell wall by means of the cell wall
recycling pathway. We show that Escherichia coli is able to import exogenous added
murein tripeptide labeled with N-7-nitro-2,1,3-benzoxadiazol-4-yl (AeK-NBD) into the
cytoplasm where it enters the peptidoglycan biosynthesis route via the tripeptide ligase
Mpl, resulting in fluorescent labels specifically located in the cell wall. When wild-type
cells were grown in the presence of the fluorescent peptide, peptidoglycan was uniformly
labeled in cells undergoing elongation. Cells in the process of division displayed a lack of
labeled peptidoglycan at mid-cell. Analysis of labeling patterns in cell division mutants
showed that the occurence of unlabeled peptidoglycan is dependent on the presence of
FtsZ, but independent of FtsQ and FtsI. Accumulation of fluorescence at division sites of a
triple amidase mutant (∆amiABC) revealed that AeK-NBD is incorporated into septal
peptidoglycan. AmiC was shown to be involved in the rapid removal of labeled
peptidoglycan side chains at division sites. As septal localization of AmiC is dependent on
FtsQ and FtsI, this demonstrates the presence of peptidoglycan hydrolase activity directly
dependent on FtsZ.
Introduction
The peptidoglycan (murein) sacculus surrounding bacterial cells is essential for their
survival and its synthesis machinery forms an important target for antibiotics. The sacculus
provides the cell with osmotic stability and it accounts for the preservation of a defined cell
shape. Furthermore, it has a vital role in bacterial morphogenesis [1-4]. It is a continuous
macromolecular structure consisting of repeating subunits of N-acetylglucosamine
(GlcNAc) and N-acetylmuramic acid (MurNAc). MurNAc residues are substituted by short
peptides, which cross-link neighboring strands to form a rigid net-like structure. In nascent
peptidoglycan the peptide part most frequently consists of L-alanyl-γ-D-glutamyl-mesodiaminopimelyl (or L-lysyl)-D-alanyl-D-alanine. The delivery of the peptidoglycan
building blocks can occur in two ways: de novo synthesis and cell wall recycling (Fig. 1).
De novo biosynthesis of peptidoglycan starts in the cytosol with the assembly of UDP34
MurNAc-pentapeptide by step-wise addition of the amino acids to UDP-MurNAc. Each
step is catalyzed by a highly specific synthetase (MurC, MurD, MurE and MurF) (Fig. 1).
On the cytoplasmic side of the plasma membrane the UDP-activated aminosugars are then
assembled on membrane-embedded bactoprenyl phosphate to form Lipid II, the final
precursor product. Lipid II is translocated to the periplasmic side of the plasma membrane
by (a) thus far unknown protein(s) [5], where it is used as a substrate for polymerization
reactions. These reactions are catalyzed by high molecular weight penicillin-binding
proteins (PBPs) or monofunctional transglycosylases [6, 7].
During growth and cell division E. coli breaks down a high fraction of its peptidoglycan by
the action of lytic transglycosylases, amidases and endopeptidases. E. coli possesses an
efficient recycling pathway of peptidoglycan building blocks released from the sacculus
[8]. This recycling pathway includes transport of the breakdown products into the cell
followed by reentering the biosynthesis pathway and reincorporation into the cell wall. The
degradation products of the lytic transglycosylases are anhydromuropeptides. These are
imported into the cytoplasm via a specific permease, AmpG, where they are further
degraded by the combined action of various enzymes. The end product of this degradation
pathway regarding the peptide part is murein tripeptide (L-alanine-γ-D-glutamine-mesodiaminopimelic acid) [9].
Alternatively, peptide turnover products are to a minor extent obtained in the periplasm by
amidases that cleave the peptide from the aminosugar parts. Murein tripeptide freed in the
periplasm this way is recycled by making use of the oligopeptide permease (Opp) pathway
[9]. This is the major pathway responsible for the uptake of small peptides in E. coli [10].
The murein tripeptide first specifically binds to the periplasmic protein MppA (murein
peptide permease A) which then utilizes membrane-bound and cytoplasmic Opp
components to transport the peptide into the cytoplasm [11]. In both pathways murein
tripeptide is directly ligated to UDP activated MurNAc by Mpl (murein peptide ligase) and
reenters the biosynthesis pathway for murein synthesis (Fig. 1) [12]. None of the gene
products participating in peptidoglycan recycling has been shown to be essential for cell
survival, except the L,D-carboxypeptidase LdcA, which is not required during exponential
growth but becomes essential when cells enter the stationary phase of growth [13].
However, AmpG and the cytoplasmic 1,6-anhydro-N-acetylmuramyl-L-alanine amidase
AmpD are essential components of the inducible ß-lactamase system of many bacteria, and
mutant strains lacking lytic transglycosylase activity fail to induce ß-lactamase [14].
35
Figure 1. Peptidoglycan recycling and de novo synthesis. Murein turnover products are taken up by
the AmpG, Mpp and Opp systems and degraded to form murein tripeptide. The muropeptide ligase
Mpl adds the tripeptide to UDP-MurNAc. With the formation of UDP-MurNAc-tripeptide, the
recycling pathway merges with de novo peptidoglycan synthesis pathway. M, N-acetylmuramic acid;
G, N-acetylglucosamine; PBP, penicillin-binding protein; A, alanine; E, glutamine; Dap,
diaminopimelic acid; anh, anhydro.
Therefore, it is thought that recycling may have a signaling function by which the cell can
monitor the condition of its cell wall structure and metabolism. For example, the rate and
mode of peptidoglycan synthesis could be mediated by variations in the level of recycled
compounds in the cytoplasm [15, 16]. The recycling pathway could also primarily function
to avoid innate immune response, by preventing the substantial release of immune response
inducing peptidoglycan degradation product.
Escherichia coli has two distinct spatial modes of peptidoglycan synthesis during cell
growth: elongational and septal. Synthesis during cell elongation involves the incorporation
36
of the disaccharide pentapeptide into the existing sacculus over the entire surface of the
cylindrical lateral wall. Cell division-associated (septal) growth is initiated after the cell has
grown to sufficient length and the DNA has been duplicated and is in the process of being
segregated into two daughter nucleoids. The cell division begins with the formation of a
septum at mid-cell and results in the constriction of the cell wall and cell membranes that
leads to the formation of two daughter cells [3, 17, 18]. Peptidoglycan formed during
septation is believed to be metabolically inert, thus forming the new cell poles in daughter
cells [19, 20]. The division process requires the concomitant action of an assembly of at
least ten cell division proteins, termed the divisome. The precise function of most of these
proteins is still unknown [21]. The assembly of these proteins at the division site occurs in a
specific and interdependent order and it was shown to occur in two steps separated by a
time-delay [22]. The first step involves the polymerization of FtsZ in a ring-like structure
and the simultaneous localization of FtsA, ZipA, ZapA and probably FtsE/X at mid cell
[22-25]. During the second step FtsK, FtsQ, FtsL, YgbQ(FtsB), FtsW, PBP3, FtsN, and
AmiC are recruited to the division site [22, 26, 27]. The function of each assembly step is
also unclear. The first step seems to be involved in the switch from cylindrical to septal cell
wall synthesis [19, 20] whereas the much later localizing cell division proteins are
responsible for the modification of the envelope shape into that of two new poles [22, 25,
28].
During cell division, enzymes capable of cleaving different bonds within the peptidoglycan
cell wall play a vital role. E. coli possesses numerous peptidoglycan hydrolases:
endopeptidases cleave cross-linked peptides that connect the glycan chains, lytic
transglycosylases cut the glycan backbone, and N-acetylmuramyl-L-alanine amidases
(AmiA, AmiB and AmiC) cleave the bond between the peptide and N-acetylmuramic acid
thereby removing peptide side chains from the carbohydrate polymer [29]. E. coli also has
two 1,6-anhydro-N-acetylmuramyl-L-alanine amidases, the aforementioned AmpD and the
recently discovered AmiD, which is also capable of cleaving the bond between MurNAc
and L-alanine but is not involved in cell separation [30]. The peptidoglycan hydrolases
participate in fundamental biological functions by allowing the insertion of newly
synthesized cell wall components into the existing peptidoglycan [31]. Moreover, E. coli
mutants lacking multiple hydrolases show defects with regard to cell separation, indicating
that these enzymes are involved in cleaving the septum to separate daughter cells during
division [32]. Judging from phenotypes of the various hydrolase mutants, the amidases (N37
acetylmuramyl-L-alanine amidases), of which only AmiC has been shown to localize to the
septal ring during division [33], appear to play the predominant part [34], although a recent
report attributed only a minor role for the amidases during cell separation [35]. E. coli
mutants lacking all three amidases are unable to separate daughter cells after division and
grow in chains of cells with thick septal peptidoglycan rings at division sites. These effects
are diminished when any one of the amidases is still active, indicating that they have
overlapping or redundant functions [28, 32]. The exact physiological functions of the
amidases and other enzymes responsible for modification and breakdown of peptidoglycan
remain poorly understood.
In this study we describe a new labeling approach, which we applied to study the cell wall
growth in vivo in E. coli, and which is potentially applicable in multiple areas of cell wall
synthesis research. This approach relies directly on the biosynthetic incorporation of
labeled cell wall precursors into the sacculus. The results presented here show that when
murein tripeptide (L-alanyl-γ-D-glutamyl-L-lysine) labeled at the lysine residue with N-7nitro-2,1,3-benzoxadiazol-4-yl (AeK-NBD) (Fig. 2) is supplied extracellularly to bacterial
cells, it is taken up by the cell and incorporated into the pre-existing sacculus via the cell
wall recycling pathway. Taking advantage of the incorporation of the fluorescent peptide
into the cell wall, we employed this method to get more insight into the spatial and
temporal relation between peptidoglycan synthesis, modification/breakdown by amidases
and the exact stage of cell growth and division. Fluorescence microscopy analysis showed
that during cell elongation the cylindrical part of the cell wall was uniformly labeled. A
decrease of labeling was observed at mid-cell in cells that were undergoing division.
Labeling patterns in cell division mutant strains showed that the lack of labeling was
dependent on the presence of FtsZ. Labeling of PG in amidase mutants revealed that the
fluorescent precursor is in fact also incorporated into septal peptidoglycan and that AmiC is
involved in the subsequent rapid removal of labeled side chains. The FtsZ dependent
removal of labels implies the presence of peptidoglycan hydrolase activity directly
activated by FtsZ during the switch from cell elongation to septum formation.
38
Figure 2. Structure of murein tripeptide (L-alanyl-γ-D-glutamyl-L-lysine) labeled at the lysine residue
with N-7-nitro-2,1,3-benzoxadiazol-4-yl (AeK-NBD).
Materials and methods
Bacterial strains, plasmids and growth conditions. The E. coli strains used in this study are
listed in Table 1. W3899 and TP980 were grown at 37°C in rich medium containing 10 g
bactotryptone, 5 g yeast extract, 15 mmol NaOH per liter and 10 g NaCl (LB). When
required, 40 µg/ml kanamycin (kan) was added to medium (Table 1). LMC500, LMC531,
LMC510, MHD44 and MHD52 cells were grown to steady state at 28 °C in glucose
minimal (GB1) medium containing 6.33 g of K2HPO4.3H2O, 2.95 g of KH2PO4, 1.05 g of
(NH4)SO4, 0.10 g of MgSO4.7H2O, 0.28 mg of FeSO4.7H2O, 7.1 mg of Ca(NO3)2.4H2O, 4
mg of thiamine, 4 g of glucose and 50 µg of required amino acids per liter pH 7.0. LMC509
was grown to steady state at 28°C in ½ GB1, because of the salt suppression of the Ts
mutation. LMC500, LMC509, LMC531 and LMC510 require lysine for growth in minimal
medium (Table 1) [36]. For MHD44 and MHD52 20 µg/ml chloramphenicol (cm) and 50
µg/ml kanamycin was added to medium, respectively. Absorbance was measured at 450 nm
(GB1), or at 600 nm (LB) with a 300-T-1 spectrophotometer (Gilford Instrument
Laboratories Inc.). Cell numbers were monitored using an electronic particle counter
39
(orifice 30 µm). Cultures were considered to be in steady state of growth if the ratio
between optical density and number of cells remained constant over time [37].
Table 1. Bacterial strains.
Strain/plasmid Relevant characteristic
Reference/source
E. coli strain
W3899
Coli Genetic
Stock Center
F-, araD139,∆(argF-lac)U169deoC1, flbB5301,
LMC500
[38]
(MC4100lysA) ptsF25, rbsR, relA1, rpslL150, lysA1
LMC509
MC4100LysA ftsZ84(Ts)
[38]
LMC531
MC4100LysA ftsQ1(Ts)
[38]
LMC510
MC4100LysA ftsI2158 (Ts)
[38]
TP980
λ- dapD2 relA1 spoT1 thi-1 mpl::kan
[12]
MHD44
MC1061 amiA, amiB::cm
[34]
MHD52
MC1061 amiA::cm amiB, amiC:: kan
[34]
Synthesis of AeK-NBD. NBD-labeled tripeptide (AeK-NBD) (Fig. 2) was synthesized on
the solid phase using tentagel S PHB resin. Fmoc-Lys(NBD)-OH was loaded on the resin
according to the method of Sieber for a loading of 0.155 mmol/g [39]. Fmoc-D-Glu(OH)OtBu,
and
Boc-Ala-OH
were
sequentially
coupled
using
(benzotriazol-1-
yloxy)tris(dimethylamino)-phosphonium hexafluorophosphate (BOP) as the coupling
reagent using standard solid phase peptide synthesis procedures. The labeled tripeptide was
cleaved from the resin with TFA, using TIS and H2O as scavengers, and purified to >98%
purity using preparative HPLC. The peptide was verified as the correct product with mass
spectrometry. Calculated mass: 509 Da. Found: [M+H]+ 509.85 Da (data not shown).
Incorporation of AeK-NBD into the bacterial cell wall of E. coli. E. coli wild-type cells
(W3899) were grown with aeration in LB broth overnight at 37°C. The overnight culture
40
was diluted 20-fold in fresh LB medium and grown with aeration to mid-log phase. Cells
were harvested by centrifugation at 2700 g for 3 min at 4°C, and washed with a buffer
containing 10 mM Tris-HCl pH 7.5, 100 mM NaCl, 2 mM MgSO4, 0.4% (w/v) glucose.
Subsequently, the cells were incubated at an OD 600 of ~1.0 in the same buffer containing
1.0 mM AeK-NBD for 1 h at 37°C. Cells were collected by centrifugation as described and
washed three times with buffer.
For steady state experiments, cells were grown to steady state as described above. AeKNBD was added to the growth medium to a final concentration of 1.0 mM and the cells
were grown for another 2-4 mass doublings. For pulse chase experiments, cells were grown
for 2 mass doublings in the presence of 1.0 mM AeK-NBD. Cells were then diluted four
fold in medium without the fluorescent peptide, grown for one mass doubling and analyzed
by fluorescence microscopy.
Isolation of peptidoglycan and fluorescence spectrometry. E. coli peptidoglycan was
isolated essentially as published with some modifications [40]. Briefly, cells were grown to
mid-log phase (OD 600 of ~0.6) and 2 ml of this culture was harvested, washed and
incubated with AeK-NBD. Cells were chilled rapidly after incubation, harvested by
centrifugation and resuspended in 1 ml of ice-cold water. The suspension was slowly
dropped into an equal volume of boiling 8% (w/v) sodium dodecyl sulfate (SDS) solution
and boiled for 30 min. After standing overnight at room temperature the sacculi were
collected by ultracentrifugation (TLA 100.1 rotor, 90.000 rpm, 45 minutes at 30°C) and
washed four times with deionized water to reduce the SDS content. For protease digestion
the sacculi were resuspended in 500 µL of 50 mM sodium phosphate buffer (pH 7.5) and
incubated with α-chymotrypsin (100 µg/ml) first for 4 h at 37 °C and then overnight after
the addition of a second equal dose of enzyme [19]. After repeating the boiling, washing
and centrifugation procedure, sacculi were finally resuspended in 125 µL of 50 mM
phosphate buffered saline (PBS) pH 7.5.
Fluorescence measurements were performed in a quartz cuvette at room temperature on a
QM-1 fluorometer (Photon Technology International Inc.). The excitation wavelength was
set to 482 nm and emission spectra were recorded between 500 and 580 nm.
Lipid II extraction. Lipid II was extracted using its unusual pH-dependant solubility
property in a Bligh and Dyer system [41, 42]. E. coli cells were grown to mid-log phase and
5 ml of this culture was harvested, washed and incubated with AeK-NBD as described
above. Cells were collected by centrifugation and washed three times as described and the
41
cell pellet was resuspended in 400 µL 50 mM PBS pH 7.5, followed by the addition of 500
µl chloroform and 1 ml methanol. After vortexing for 10 min at room temperature, the
homogenates were centrifuged at 4000g for 5 min and the supernatants were converted to a
two-phase system by adding 500 µL chloroform and 500 µl PBS. The aqueous methanol
(upper) phase was acidified to pH 1 by the addition of concentrated HCl (37%).
Subsequently, 500 µl of a fresh lower phase derived from a pre-equilibrated two-phase
acidic Bligh–Dyer system was added. After thorough mixing, the phases were separated by
centrifugation at 4000g for 2 min and the lower phase was dried in a vacuum dessicator.
The lipid film was redissolved in 20 µL chloroform/methanol (1:1 v/v) and analyzed by
TLC using chloroform/methanol/water/ammonia (88:48:10:1) [43]. Spots were visualized
by UV illumination and by staining with iodine vapor.
Microscopy and image analysis. Objects containing immobilized E. coli were analyzed on a
Nikon Eclipse TE2000 inverted microscope with a Pan Fluor 60.0x/1.40/0.21 oil objective
was used. NBD was excited by an Argon laser (488 nm, Spectra-Physics) and Nikon EZ-C1
Software Version 2.20 Gold was used for analysis of the images. Difference Interference
Contrast (DIC) (Nomarsky optics) was used for detection of bacteria without a fluorescent
signal or for comparison with fluorescent cells. Fluorescence microscopy and image
analysis on steady state grown cells and temperature-sensitive cell division mutants were
performed as previously described [22].
Results
Incorporation of AeK-NBD into the cell wall via the recycling pathway. To incorporate
fluorescent labels into the growing peptidoglycan layer of living bacteria, cells were
incubated with AeK-NBD and subsequently analyzed by confocal fluorescence
microscopy. Brightly fluorescent cells were observed as is shown in an overview (Figure
3A). Comparison between images at higher magnification in differential interference
contrast (DIC) (Fig. 3B) and fluorescence mode (Fig. 3C) showed that the peptide had been
taken up by nearly all the cells.
42
Figure 3. Overview of E. coli W3899 cells incubated with AeK-NBD as described in ‘Experimental
procedures’ (A). Image of the cells at higher magnification in DIC (B) and fluorescence (C) mode.
To get insight into whether or not the fluorescent peptide was incorporated into the
peptidoglycan, sacculi were isolated from the same number of cells that were incubated
with the fluorescent peptide and analyzed by fluorescence spectroscopy. The results of a
typical experiment are shown in Figure 4. The isolated sacculi are clearly fluorescently
labeled following incubation with AeK-NBD (Fig. 4, spectrum 1). Subsequent protease
treatment followed by washing did not significantly decrease the fluorescence intensity but
resulted in a small red-shift (data not shown), indicating a more hydrophilic
microenvironment, suggesting that some labels had an interaction with one or more
proteins.
In principle, it could be possible that the tripeptide is somehow directly incorporated into
the cell wall without first entering the recycling route. To exclude this possibility, we tested
the incorporation efficiency of the labeled tripeptide in an Mpl deficient strain (TP980).
Mpl (murein peptide ligase) attaches recycled murein tripeptide directly to UDP-MurNAc,
which is the final step after which the peptide re-enters the cell wall biosynthesis cycle [12].
Therefore, this mutant is unable to synthesize peptidoglycan via the recycling pathway. Mpl
deficient cells were incubated with AeK-NBD and sacculi were isolated and analyzed for
43
the presence of fluorescent label. Additionally, cells were analyzed by confocal
fluorescence microscopy. Fluorescence intensity measurements showed that no label had
been incorporated into the sacculi (Fig. 4, spectrum 3). Moreover, cells were hardly
distinguishable from background on fluorescence microscopy images (Fig. 5A), even
though they still possess the proteins necessary for the uptake of the tripeptide. This
indicates that unprocessed AeK-NBD residing in the cytoplasm contributes only to a minor
extent to the fluorescent signal observed in wild-type cells treated with the fluorescent
peptide. It also points to a coupling and/or regulation between the uptake of cell wall
peptides and their utilization for synthesis in the recycling pathway.
6000
Fluorescence (a.u.)
5000
4000
1
3000
2000
3
1000
2
0
480
500
520
540
560
580
600
Wavelength (nm)
Figure 4. Fluorescence spectra of similar amounts of isolated sacculi. Wild-type E. coli W3899
incubated with AeK-NBD (1), cells grown in the absence of the peptide (2) and Mpl deficient strain
TP980 (3).
44
Figure 5. Fluorescence (A) and DIC (B) images of Mpl deficient cells (TP980) after incubation with
AeK-NBD.
Further support for the entry of the tripeptide into the bacterial cytoplasm comes from the
analysis of Lipid II in wild-type cells. Lipid II is the final membrane-bound cell wall
precursor and is present in minor quantities in bacterial cell membranes [44, 45]. If the
fluorescent tripeptide is utilized in peptidoglycan biosynthesis by means of the recycling
pathway, the cells should also contain Lipid II that carries a fluorescent label. To test this,
lipids were extracted from cells labeled with AeK-NBD using acidic Bligh and Dyer [41]
and analyzed by thin-layer chromatography (TLC). UV detection allowed for the screening
of fluorescent compounds (Figure 6). Two fluorescent compounds were observed in the
extract (Fig. 6, lane 2). Comparison with Lipid I-NBD reference standard (Fig. 6, lane 1)
clearly showed that NBD-labeled Lipid I and Lipid II, the latter which characteristically
runs at a slightly lower retention time due to the additional GlcNAc moiety [43], were
present in cells incubated with AeK-NBD. From these experiments we can conclude that
exogenously supplied AeK-NBD is taken up by the cells into the cytoplasm, where it gets
attached to UDP-MurNAc by Mpl, according to the cell wall recycling pathway. NBDlabeled Lipid II is subsequently synthesized, after which incorporation into the cell wall
takes place.
The new labeling method was then employed to study the spatial location of the products of
the recycling pathway in relation to cell growth or division. Wild-type cells of strain
LMC500, grown to steady state in GB1 at 28°C, were incubated with 1 mM AeK-NBD and
subsequently studied by fluorescence microscopy. Cells grew normally in the presence of
AeK-NBD up to four generations. Prolonged growth was somewhat inhibited probably
because the presence of the fluorescent label at the lysine residue impairs the formation of
peptide cross-links between cell wall subunits. For this reason we restricted the experiments
45
Figure 6. TLC analysis with UV detection of Lipid I and II extracted from wild-type E. coli W3899
cells incubated with AeK-NBD (lane 2). Lane 1 contains NBD labeled Lipid I as a reference.
to cells grown on the fluorescent peptide up to four mass doublings, where the presence of
the label had no effects on cell growth.
Figure 7 shows an overview of cells in different stages of growth. Distinctive labeling
patterns were observed. Elongating cells were predominately labeled uniformly with NBD.
Strikingly, cells in the process of division seemed to show a decrease of fluorescence at the
septum (Fig. 7A and C). By comparing the average fluorescence intensity profile of 103
dividing cells (Fig. 7D, grey trace) with that of the total amount of 410 cells (Fig. 7D, black
trace), a significant decrease in fluorescence at mid-cell could be observed. Figure 7A and
C also show some cells with weaker fluorescence in one of their polar caps. This is
consistent with the observation of lack of fluorescence at the cell septum. Presumably these
are recently formed daughter cells since following division, septal peptidoglycan remains at
one of the cell poles as an inert part of the sacculus [19].
46
Figure 7. Overview of steady state grown LMC500 cells labeled with AeK-NBD (A). Enlargement of
wild type cells labeled with NBD, Phase contrast (B) and fluorescence (C) image of a dividing cell
and a newly formed daughter cell showing reduced amount of fluorescence in the septum and one
polar cap respectively. The bar is 1 µm. (D) Fluorescence intensity profiles of LMC500 cells grown in
the presence of AeK-NBD for four mass doublings. The average normalized cell length is given on
the x-axis. The average normalized fluorescence intensity (a.u.) is given on the y-axis. Fluorescence
intensity was determined in 103 dividing cells (grey) and all 410 cells counted (black).
Divisome maturation and hydrolase activity. To get a more clear picture of the lack of
labeling at division sites, we studied the incorporation of the fluorescent precursor probe in
FtsZ (LMC509), FtsQ (LMC531) and FtsI (PBP3) (LMC510) temperature sensitive cell
division mutants. In the wild-type situation, FtsZ is the first protein to assemble at the site
of division, while FtsQ and PBP3 are part of the group of proteins downstream of FtsE/X
that were shown to assemble in the second step. The mutant strains are isogenic to LMC500
47
apart from the Ts mutation (see Table 1). The temperature sensitive strains were first grown
to steady state at 28°C. The cultures were then shifted to 42°C for 4 mass doublings in the
presence of 1 mM AeK-NBD. FtsQ filaments show aborted so called blunt constrictions as
previously described (Fig. 8A-D) [38]. After 2 mass doublings, these filaments show a
regular fluorescence staining of the peptidoglycan apart from the two most recently
synthesized blunt constrictions, which are clearly non-fluorescent (Fig. 8B). The central
blunt constriction, however, is labeled with the precursor probe. After 4 mass doublings
again non-fluorescent bands are visible at the new potential division sites (Fig. 8D). The
previously unlabeled division sites from the preceding generation appear to have
incorporated some fluorescent precursors. Subsequent labeling experiments in an FtsI
(PBP3) Ts strain gave similar labeling patterns as in the FtsQ mutant (data not shown). The
FtsZ filaments are smooth with sometimes a hint of a constriction (Fig. 8E) and show a
regular staining of the peptidoglycan layer apart from now and then a very narrow dark
band perpendicular to the length axes of the bacterium (Fig. 8F). In contrast, the nature of
occurrence of several small, unlabeled bands in the FtsZ deficiency strain is not completely
clear, but it may be attributed to slight leakiness of the mutant. This shows that the lack of
fluorescence at division sites is dependent on the presence of FtsZ, but independent of FtsQ
and FtsI.
Pulse-chase experiments revealed that the fluorescent labels had disappeared after a chase
period of one generation time (data not shown). To investigate whether the lack of labeling
at the division site is caused by a specific removal of the labeled murein peptides by
amidases, we studied the incorporation of the fluorescent precursor probe in two amidase
mutants; the triple mutant MHD52 (∆amiABC) and the double mutant MHD44 (∆amiAB).
Mutants lacking two or more amidases are unable to divide and grow in long chains of
connected cells with thick septal peptidoglycan rings at division sites [28, 32]. MHD52 and
MHD44 were incubated with 0.2 mM AeK-NBD and fluorescence was monitored in time.
Almost all cells of the triple amidase mutant showed increased fluorescence at incipient
septa near the end of one growth cycle (Fig. 9). This demonstrates that the fluorescent
precursors are also incorporated into the septal peptidoglycan and that they are not (or very
slowly) removed in the absence of the amidases. Increased fluorescence at the septa was not
observed in the double mutant MHD44 (Figure 10), demonstrating that AmiC is involved in
the removal of labeled side chains of septal peptidoglycan.
48
Figure 8. A: Phase contrast (A and C) and fluorescence images (B and D) of LMC531(FtsQ1Ts)
grown for 2 and 4 mass doublings, respectively, at the restrictive temperature in 1/2GB1 in the
presence of AeK-NBD. Phase contrast (E) and fluorescence image (F) of LMC509(FtsZ84Ts) grown
for 4 mass doublings at the restrictive temperature in GB1 in the presence of AeK-NBD.
49
Thus, the lack of fluorescence at the division sites in wild-type cells can be attributed in
large part to the elevated activity of AmiC during septal peptidoglycan synthesis. As septal
localization of AmiC is dependent on the presence of divisome components including FtsQ
and FtsI [33], our data imply that the localized activity of a yet undetermined PG
hydrolase(s) is directly triggered upon FtsZ assembly at (future) division sites.
Figure 9. Phase contrast (A) and fluorescence image (B) of MHD52 (∆amiABC) grown for nearly 1
mass doubling in GB1 in the presence of AeK-NBD.
50
Figure 10. Phase contrast (left) and fluorescence image (right) of MHD44(∆amiAB) grown for 2 mass
doublings at 28°C in GB1 in the presence of AeK-NBD.
Discussion
In this paper, we have described a new labeling approach for the study of peptidoglycan
synthesis and turnover in relation to cell growth and division in E. coli. The method
involves supplying cells with the cell wall precursor murein tripeptide (AeK) fluorescently
labeled with NBD. Thus, the label enters the cell wall biosynthesis pathway almost at the
earliest possible time point. This method of incorporation of reporter groups in the cell wall
of bacteria does not require (mis)treatment of the cells in any way and it complements a
previously described labeling method, which was based on amino acid exchange by yet
uncharacterized enzymatic events in pre-existing peptidoglycan [19, 46]. Moreover, by
directly utilizing the cell wall biosynthesis machinery to achieve labeling, we were able to
shine new light on the synthesis and turnover of polar peptidoglycan.
Our results show that in vivo AeK-NBD is indeed efficiently incorporated into the cell wall
by making use of the cell wall recycling pathway. In order to attribute the observed
fluorescence patterns, labeled molecules present in the cell can be categorized into three
groups: soluble cell wall precursors, mature cell wall and Lipid II in the cytoplasmic
membrane. Lipid II is known to be present in the cell only in very low amounts [44, 45].
Moreover, Errington et al. established that in rod-shaped bacteria like B. subtilis and E.
51
coli, Lipid II has a septal localization during cell division and a helical localization along
the length of the cell during elongation [2], which is unlike the labeling patterns we
observe. The lack of contribution to the fluorescence by soluble precursors became
apparent with the occurrence of dark bands at potential division sites in FtsQ filaments. A
local delocalization of cytosolic cell wall precursors would be highly improbable. Also,
hardly any fluorescence was observed in the Mpl mutant, where the fluorescent tripeptide
could not be utilized for cell wall synthesis. These observations suggest that the vast
majority of the observed fluorescence can be attributed to labels residing in the cell wall.
The most striking observation following analysis of the labeling patterns is the lack of
fluorescent labels in the cell wall at division sites. In theory, this phenomenon could be
explained in two ways. One is the rapid breakdown of labels after synthesis of septal
peptidoglycan and another explanation could be the exclusion of recycled material during
septal synthesis. It has been shown that a high rate of PG degradation occurs during
septation. Over 30% of newly synthesized septal peptidoglycan is rapidly turned over by
the combined action of multiple hydrolases [35]. Our labeling experiments in the triple
amidase mutant showed the opposite effect as compared to wild-type cells, i.e. increased
fluorescence at division sites, whereas this effect was diminished in the double mutant
where only AmiC is active. As it is known that thick rings of septal peptidoglycan are
present in chain-forming amidase mutants [34], this clearly demonstrates that labeled
precursors are utilized for septal peptidoglycan synthesis and that AmiC activity is mainly
responsible for the rapid removal of labeled murein peptides at division sites in wild-type
cells. It also shows that in the absence of amidases, other peptidoglycan hydrolases that
may still be present provide an insignificant or no contribution at all to the removal of
labeled peptides from septal peptidoglycan subunits. The presence of bands of unlabeled
material in the division mutants FtsQ(Ts) and FtsI(Ts) at the restricted temperature suggests
a highly localized activity of peptidoglycan hydrolases. Initiation of hydrolase activity
seems to be dependent on the presence of FtsZ. It was shown in C. crescentus that FtsZ
plays a major role in the recruitment of Lipid II synthesizing enzymes to mid-cell before
constriction and the so called preseptal elongation, which occurs in a growth zone at midcell, whereas constriction starts with the assembly of the second step division proteins [27].
It has been proposed that FtsZ could have similar functions in other bacterial species
including E. coli [20, 27, 47].
52
Our results indicate that E. coli indeed possesses a preseptal FtsZ-dependent synthesis to
which apparently PG hydrolase activity is recruited, which explains the dark bands in
filamentous FtsQ(Ts) and FtsI(Ts) cells. AmiC, which is involved in septal peptidoglycan
synthesis and daughter cell separation, is the only amidase shown to be a specific
component of the septal ring [28, 33], although preliminary studies have recently been
mentioned which suggest the vast enrichment of AmiB at the division site [48]. The
localization of AmiC to the division site is dependent on the preassembly of division
proteins, including FtsQ [28, 33]. This implies that AmiC is not localized at the division
sites in the FtsQ(Ts) filaments and that AmiA, AmiB and/or other PG hydrolases are
responsible for the observed activity during preseptal PG synthesis. There are several
previous observations which point to the possibility of PG hydrolase activity during FtsZdependent elongation. FtsI(Ts) filaments treated with the β-lactam cefsulodin lysed faster
than FtsZ(Ts) filaments and sacculi displayed extremely sharp transverse cuts at potential
division sites [49]. This suggests that inhibition of the synthetic enzymes in FtsZ-associated
biosynthetic complexes leads to rapid PG degradation by the action of hydrolases.
Furthermore, depletion of FtsZ in E. coli strains lacking PBP5, a D,D-carboxypeptidase
responsible for removing the terminal D-Ala from pentapeptides, resulted in a reduced
precursor incorporation into the side walls near the poles, together with a significant
increase in the pentapeptide levels [20]. This indicates that FtsZ-directed sidewall synthesis
utilizes pentapeptide-containing muropeptides. In possible relation to this, AmiA and AmiB
were shown to have a preference for removing pentapeptide side chains, as AmiA or AmiB
deletion in PBP5 mutants resulted in a significant increase in the level of pentapeptides [34,
50]. Therefore, it is conceivable that AmiA and/or AmiB activity contributes to the
pentapeptide
consumption
during
FtsZ-mediated
synthesis.
Also,
the
putative
endopeptidase EnvC was shown to localize to the nascent division site prior to
invagination, and was seemingly not required throughout septation, indicating that it acts at
a very early stage of the division process [28, 51].
FtsZ-directed preseptal synthesis results in a ring of inert peptidoglycan [19], which could
be the earliest differentiation stage of the division site. We can envision a role for PG
hydrolases in modifying the three-dimensional PG structure during preseptal synthesis, e.g.
regulation of the number of cross-links, to eventually produce an inert PG ring with the
correct geometry for ensuing septum formation. This would be similar to events that occur
53
during septal PG synthesis, where amidases and other hydrolases must maintain the correct
balance between inert PG formation and PG insertion at the leading edge.
Interestingly, in the FtsQ mutant we observed that the blunt constriction at mid cell did
contain a considerable amount of labels. This can be explained by the fact that FtsZ initially
localized to initiate division but that the Z-ring was unstable and disassembled [52], thereby
causing the delocalization of the amidase activity, to subsequently assemble at one quarter
and three quarter positions. The disassembly probably did not give the enzymes involved in
FtsZ-directed synthesis sufficient time to modify the PG into a completely inert PG ring,
thus insertion of labeled PG continued at the (former) division sites following Z-ring
disassembly.
Whether removal of labeled peptidoglycan subunits is a general consequence of the high
activity of amidases during cell division or is specific for the presence of labels is unclear.
Pulse-chase experiments revealed that the majority of labels present in the cell wall were
rapidly removed in the presence of amidases as soon as cells were transferred into chase
medium. This suggests that in the continuous presence of the fluorescent peptide, the
amidase activity during cell elongation is not sufficient to keep up with the rate of
incorporation of labeled precursors. The specific muropeptide composition has been shown
to be an important factor in the spatial organization of peptidoglycan synthesis during
different stages of the cell cycle [20, 53]. Also, subunits containing the NBD-label are
impaired in peptide cross-linking, which could weaken the cell wall in this proximity. This
could serve as a recognition point for the amidases that modify the peptidoglycan. It may
even be an indication for the existence of a peptidoglycan repair mechanism, similar to that
of damaged DNA.
Finally, our novel in vivo cell wall labeling method also opens up the possibility for a
proteomics approach. Our next objective is to incorporate cell wall precursors labeled with
photo-activatable crosslinkers, in order to identify and study proteins interacting with these
precursors and the peptidoglycan cell wall. This could lead to the discovery of new targets
for antibiotics.
Acknowledgments
This work was supported by a Vernieuwingsimpuls grant from the Netherlands
Organization for Scientific Research (NWO). Jolanda Verheul is acknowledged for
technical assistance.
54
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
van Heijenoort, J., Recent advances in the formation of the bacterial
peptidoglycan monomer unit. Nat Prod Rep, 2001. 18(5): p. 503-19.
Daniel, R.A. and J. Errington, Control of cell morphogenesis in bacteria: two
distinct ways to make a rod-shaped cell. Cell, 2003. 113(6): p. 767-76.
den Blaauwen, T., et al., Morphogenesis of rod-shaped sacculi. FEMS Microbiol
Rev, 2008. 32(2): p. 321-44.
Vollmer, W., D. Blanot, and M.A. de Pedro, Peptidoglycan structure and
architecture. FEMS Microbiol Rev, 2008. 32(2): p. 149-67.
van Dam, V., et al., Transmembrane transport of peptidoglycan precursors across
model and bacterial membranes. Mol Microbiol, 2007. 64(4): p. 1105-14.
van Heijenoort, J., Formation of the glycan chains in the synthesis of bacterial
peptidoglycan. Glycobiology, 2001. 11(3): p. 25R-36R.
Sauvage, E., et al., The penicillin-binding proteins: structure and role in
peptidoglycan biosynthesis. FEMS Microbiol Rev, 2008. 32(2): p. 234-58.
Park, J.T. and T. Uehara, How bacteria consume their own exoskeletons (turnover
and recycling of cell wall peptidoglycan). Microbiol Mol Biol Rev, 2008. 72(2): p.
211-27, table of contents.
Goodell, E.W. and C.F. Higgins, Uptake of cell wall peptides by Salmonella
typhimurium and Escherichia coli. J Bacteriol, 1987. 169(8): p. 3861-5.
Higgins, C.F. and M.M. Gibson, Peptide transport in bacteria. Methods Enzymol,
1986. 125: p. 365-77.
Park, J.T., et al., MppA, a periplasmic binding protein essential for import of the
bacterial cell wall peptide L-alanyl-gamma-D-glutamyl-meso-diaminopimelate. J
Bacteriol, 1998. 180(5): p. 1215-23.
Mengin-Lecreulx, D., J. van Heijenoort, and J.T. Park, Identification of the mpl
gene encoding UDP-N-acetylmuramate: L-alanyl-gamma-D-glutamyl-mesodiaminopimelate ligase in Escherichia coli and its role in recycling of cell wall
peptidoglycan. J Bacteriol, 1996. 178(18): p. 5347-52.
Templin, M.F., A. Ursinus, and J.V. Holtje, A defect in cell wall recycling triggers
autolysis during the stationary growth phase of Escherichia coli. Embo J, 1999.
18(15): p. 4108-17.
Korsak, D., S. Liebscher, and W. Vollmer, Susceptibility to antibiotics and betalactamase induction in murein hydrolase mutants of Escherichia coli. Antimicrob
Agents Chemother, 2005. 49(4): p. 1404-9.
Jacobs, C., et al., Bacterial cell wall recycling provides cytosolic muropeptides as
effectors for beta-lactamase induction. Embo J, 1994. 13(19): p. 4684-94.
Park, J.T., Why does Escherichia coli recycle its cell wall peptides? Mol
Microbiol, 1995. 17(3): p. 421-6.
Canepari, P., et al., Cell elongation and septation are two mutually exclusive
processes in Escherichia coli. Arch Microbiol, 1997. 168(2): p. 152-9.
Nanninga, N., Morphogenesis of Escherichia coli. Microbiol Mol Biol Rev, 1998.
62(1): p. 110-29.
de Pedro, M.A., et al., Murein segregation in Escherichia coli. J Bacteriol, 1997.
179(9): p. 2823-34.
55
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
Varma, A., M.A. de Pedro, and K.D. Young, FtsZ directs a second mode of
peptidoglycan synthesis in Escherichia coli. J Bacteriol, 2007. 189(15): p. 5692704.
Errington, J., R.A. Daniel, and D.J. Scheffers, Cytokinesis in bacteria. Microbiol
Mol Biol Rev, 2003. 67(1): p. 52-65.
Aarsman, M.E., et al., Maturation of the Escherichia coli divisome occurs in two
steps. Mol Microbiol, 2005. 55(6): p. 1631-45.
Rueda, S., M. Vicente, and J. Mingorance, Concentration and assembly of the
division ring proteins FtsZ, FtsA, and ZipA during the Escherichia coli cell cycle.
J Bacteriol, 2003. 185(11): p. 3344-51.
Corbin, B.D., et al., Interaction between cell division proteins FtsE and FtsZ. J
Bacteriol, 2007. 189(8): p. 3026-35.
Den Blaauwen, T., et al., Timing of FtsZ assembly in Escherichia coli. J Bacteriol,
1999. 181(17): p. 5167-75.
Wang, X., C. Possoz, and D.J. Sherratt, Dancing around the divisome: asymmetric
chromosome segregation in Escherichia coli. Genes Dev, 2005. 19(19): p. 236777.
Aaron, M., et al., The tubulin homologue FtsZ contributes to cell elongation by
guiding cell wall precursor synthesis in Caulobacter crescentus. Mol Microbiol,
2007. 64(4): p. 938-52.
Priyadarshini, R., M.A. de Pedro, and K.D. Young, Role of peptidoglycan
amidases in the development and morphology of the division septum in
Escherichia coli. J Bacteriol, 2007. 189(14): p. 5334-47.
Holtje, J.V., Growth of the stress-bearing and shape-maintaining murein sacculus
of Escherichia coli. Microbiol Mol Biol Rev, 1998. 62(1): p. 181-203.
Uehara, T. and J.T. Park, An anhydro-N-acetylmuramyl-L-alanine amidase with
broad specificity tethered to the outer membrane of Escherichia coli. J Bacteriol,
2007. 189(15): p. 5634-41.
Holtje, J.V., From growth to autolysis: the murein hydrolases in Escherichia coli.
Arch Microbiol, 1995. 164(4): p. 243-54.
Heidrich, C., et al., Effects of multiple deletions of murein hydrolases on viability,
septum cleavage, and sensitivity to large toxic molecules in Escherichia coli. J
Bacteriol, 2002. 184(22): p. 6093-9.
Bernhardt, T.G. and P.A. de Boer, The Escherichia coli amidase AmiC is a
periplasmic septal ring component exported via the twin-arginine transport
pathway. Mol Microbiol, 2003. 48(5): p. 1171-82.
Heidrich, C., et al., Involvement of N-acetylmuramyl-L-alanine amidases in cell
separation and antibiotic-induced autolysis of Escherichia coli. Mol Microbiol,
2001. 41(1): p. 167-78.
Uehara, T. and J.T. Park, Growth of Escherichia coli: significance of
peptidoglycan degradation during elongation and septation. J Bacteriol, 2008.
190(11): p. 3914-22.
Koppelman, C.M., et al., R174 of Escherichia coli FtsZ is involved in membrane
interaction and protofilament bundling, and is essential for cell division. Mol
Microbiol, 2004. 51(3): p. 645-57.
Fishov, I., A. Zaritsky, and N.B. Grover, On microbial states of growth. Mol
Microbiol, 1995. 15(5): p. 789-94.
56
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
Taschner, P.E., et al., Division behavior and shape changes in isogenic ftsZ, ftsQ,
ftsA, pbpB, and ftsE cell division mutants of Escherichia coli during temperature
shift experiments. J Bacteriol, 1988. 170(4): p. 1533-40.
Sieber, P., An improved method for anchoring of 9-fluorenylmethoxycarbonylamino acids to 4-alkoxybenzyl alcohol resins. Tetrahedron Letters, 1987. 28(49):
p. 6147.
Glauner, B., Separation and quantification of muropeptides with highperformance liquid chromatography. Anal Biochem, 1988. 172(2): p. 451-64.
Guan, Z., S.D. Breazeale, and C.R. Raetz, Extraction and identification by mass
spectrometry of undecaprenyl diphosphate-MurNAc-pentapeptide-GlcNAc from
Escherichia coli. Anal Biochem, 2005. 345(2): p. 336-9.
Bligh, E.G. and W.J. Dyer, A rapid method of total lipid extraction and
purification. Can J Biochem Physiol, 1959. 37(8): p. 911-7.
Breukink, E., et al., Lipid II is an intrinsic component of the pore induced by nisin
in bacterial membranes. J Biol Chem, 2003. 278(22): p. 19898-903.
Kramer, N.E., et al., Resistance of Gram-positive bacteria to nisin is not
determined by lipid II levels. FEMS Microbiol Lett, 2004. 239(1): p. 157-61.
van Heijenoort, J., Lipid intermediates in the biosynthesis of bacterial
peptidoglycan. Microbiol Mol Biol Rev, 2007. 71(4): p. 620-35.
Vollmer, W. and J.V. Holtje, Morphogenesis of Escherichia coli. Curr Opin
Microbiol, 2001. 4(6): p. 625-33.
Vollmer, W. and U. Bertsche, Murein (peptidoglycan) structure, architecture and
biosynthesis in Escherichia coli. Biochim Biophys Acta, 2008. 1778(9): p. 171434.
Uehara, T., T. Dinh, and T.G. Bernhardt, LytM-domain factors are required for
daughter cell separation and rapid ampicillin-induced lysis in Escherichia coli. J
Bacteriol, 2009.
de Pedro, M.A., J.V. Holtje, and H. Schwarz, Fast lysis of Escherichia coli
filament cells requires differentiation of potential division sites. Microbiology,
2002. 148(Pt 1): p. 79-86.
Priyadarshini, R., D.L. Popham, and K.D. Young, Daughter cell separation by
penicillin-binding proteins and peptidoglycan amidases in Escherichia coli. J
Bacteriol, 2006. 188(15): p. 5345-55.
Bernhardt, T.G. and P.A. de Boer, Screening for synthetic lethal mutants in
Escherichia coli and identification of EnvC (YibP) as a periplasmic septal ring
factor with murein hydrolase activity. Mol Microbiol, 2004. 52(5): p. 1255-69.
Addinall, S.G., E. Bi, and J. Lutkenhaus, FtsZ ring formation in fts mutants. J
Bacteriol, 1996. 178(13): p. 3877-84.
Young, K.D., Bacterial shape. Mol Microbiol, 2003. 49(3): p. 571-80.
57
58
Chapter 3
Development of a novel proteomics
approach for the identification of
proteins interacting in vivo with the
bacterial cell wall.
Abstract
The peptidoglycan layer, cell wall, of bacteria is vital to their survival. Its metabolism is
tightly coordinated with other fundamental cellular processes and involves a vast and
complex network of enzymes. The comprehensive identification of these proteins could
provide new targets for antibiotics. For this purpose, a photo-crosslink approach was
developed to capture proteins interacting with the peptidoglycan cell wall and its metabolic
precursors or degradation products in vivo. We designed a photoactivatable analogue of
murein tripeptide containing an additional alkyne functionality (photo-AeK-alkyne). The
derivatized tripeptide enters the cell and is utilized for peptidoglycan biosynthesis via the
cell wall recycling pathway. Following UV activation, the alkyne moiety is exploited for
the selective modification of the crosslinked proteins with azide-derivatized compounds by
means of click chemistry. Fluorescent detection was enabled by the attachment of
Alexa488-azide. Initially, biotin-azide was conjugated to the proteins for affinity
purification purposes. Mass spectrometry identification revealed the presence of significant
quantities of non-specifically purified proteins. This prompted us to develop a purification
scheme that involves ‘clicking’ the crosslink products directly onto azide-derivatized
magnetic beads. Besides their intrinsically low background binding, the covalent
attachment of the proteins to the beads allowed harsh washing conditions, which were
followed by on-bead proteolytic digestion. Numerous interesting enzymes were identified
of which some were previously only hypothetically involved with cell wall metabolism.
Introduction
The peptidoglycan sacculus surrounding bacteria is essential for bacterial cell survival and
forms an important target for antibiotics. Its synthesis in bacteria requires the concerted
action of a dynamic and transiently formed network of proteins including inner membrane,
periplasmic and even outer membrane proteins in Gram-negative bacteria. Besides,
peptidoglycan synthesis is tightly linked to cell division, as some cell wall synthesizing
enzymes have been proven to be part of the divisome, an assembly of proteins that governs
cell division [1, 2]. Furthermore, enzymes interacting with peptidoglycan play important
roles in other processes, such as pathogenesis, secretion and permeation [3-5]. The exact
physiological functions of many proteins involved in peptidoglycan metabolism are still
60
unclear. Moreover, several of these enzymatic processes have not yet been unambiguously
attributed to specific enzymes, which will most likely include several hitherto unknown
proteins. In the quest for new targets for antibiotics it is therefore essential to vastly
increase our knowledge on proteins involved in bacterial cell wall metabolism.
An established method to study interactions involving proteins is photoactivatable
crosslinking combined with mass spectrometry. The interactions within a cell are dynamic
and often stabilized by one or more other proteins as part of a complex. The possibility to
comprehensively search for proteins interacting with the cell wall in the native cellular
environment would therefore be a huge benefit. In recent years, various technologies for in
vivo protein crosslinking have been developed, most notably the genetic incorporation of
non-natural, photoreactive amino acids into the primary sequence of proteins in living cells
[6]. In a different approach, photoactivatable crosslinking carbohydrates were designed and
shown to be metabolically incorporated onto the cell surface, which was utilized for
studying glycoprotein interactions [7].
For the purpose of detection and enrichment in in vivo interaction studies, bioconjugation
via copper(I)-catalyzed 1,3-dipolar cycloaddition of azides and terminal alkynes (click
chemistry) has proven to be very practical [8]. The presence of the small, physiologically
stable and inert functional groups eliminates the need for bulky reporter tags that might
unfavorably affect interactions. This was first exemplified in activity-based protein
profiling, which used active site-directed chemical probes derivatized with a ‘click’ moiety.
The combination of photo-crosslinking and click chemistry has been applied for profiling
metalloprotease inhibitors in vitro [9] and to study in situ protein-lipid and protein-small
molecule interactions [10, 11].
In this study, we describe a cell wall proteomics approach that combines in vivo photocrosslinking with click chemistry. It is based on our previously reported method to
incorporate labeled precursors into the bacterial cell wall of growing cells [12]. We had
shown that fluorescently labeled murein tripeptide (L-alanyl-γ-D-glutamyl-L-lysine) is
taken up by E. coli and utilized for peptidoglycan biosynthesis taking advantage of the
bacterial cell’s own cell wall recycling pathway (Fig. 1). We adapted this cell wall labeling
technique into a proteomics format by using a murein tripeptide analogue derivatized with
both a photoactivatable crosslinker and a ‘clickable’ alkyne moiety. Enzymes along the
metabolic pathway of peptidoglycan were shown to be crosslinked in vivo and the alkyne
functionality was utilized for detection and purification of the crosslink products through in
61
vitro conjugation of azide-modified reporter tags. In addition, a single-step purification
protocol was developed, which entails the capture of the crosslinked proteins onto azidefunctionalized magnetic beads. It directly utilizes the high specificity of the Cu(I)-catalyzed
click reaction and as an extra benefit was shown to minimize non-specific protein binding
during purification. LC-MS/MS analysis resulted in the identification of a number of
functionally interesting proteins possibly involved in peptidoglycan metabolism,
demonstrating the potential of the in vivo crosslinking approach for proteomic analysis of
the cell wall.
labeled tripeptide
PBP’s
out
Lipid I
Lipid II
Opp/MppA
MurG
UDP
MraY
G
UDP
M
A
A A
MurF
UDP
in
E K A A
M
A
E
K
Mpl
A
UDP
E
K
M
Figure 1. Biosynthetic incorporation of derivatized murein tripeptide into peptidoglycan.
62
Materials and methods
Synthesis of photo-AeK-alkyne
Photo-AeK-alkyne was a prepared by Dr. N.I. Martin, Utrecht University. The peptide was
synthesized using standard solution phase peptide synthesis protocols and reagents. The
identity of the peptide was verified using (ESI) mass spectrometry and 1H- and
13
C-NMR
and the purity was >98% as determined by TLC and NMR. Peptide was stored protected
from light at -20 °C and dissolved in the appropriate buffer immediately before use. Full
details of synthesis are to be published.
Extraction and detection of labeled Lipid II
E. coli wild type cells (W3899) were grown in LB broth overnight with aeration. The
overnight culture was diluted 20-fold in fresh LB medium and grown to mid-log phase.
Cells were harvested from 10 ml of culture by centrifugation at 2700 g for 3 min at 4°C,
and washed with a buffer containing 50 mM PBS pH 7.5, 100 mM NaCl, 2 mM MgSO4,
0.4% (w/v) glucose. Subsequently, the cells were incubated in the dark at an OD 600 of
~1.0 in 6 ml of the same buffer containing 1 mM photo-AeK-alkyne for 1 hour at 37°C.
Cells were rapidly cooled in ice-water, collected by centrifugation at 4°C, washed twice
and resuspended in 1 ml of incubation buffer. The extraction of Lipid II by acidic Bligh and
Dyer was performed essentially as described [13]. Extracts were dried under vacuum and
resuspended in 100 µl 50 mM Tris-HCl, pH 8.0 containing 1% (w/v) Triton X-100. To
fluorescently label alkyne-containing Lipid II, 1 mM TCEP, 1 mM CuSO4, 200 µM tris[(1benzyl-1H-1,2,3-triazol-4-yl)methyl]amine (TBTA) and 50 µM Alexa488-azide were
respectively added. Following incubation at RT for 2 hrs, the reaction mixture was
extracted with butanol/6M pyridine acetate, pH 4.2, after which the butanol phase was
washed with water and evaporated under vacuum. The dried extract was dissolved in
chloroform/methanol
(1:1)
and
analyzed
by
TLC
using
a
chloroform/methanol/water/ammonia (88:48:10:1) mobile phase. Detection was by UV
illumination and by staining with iodine vapor.
Extraction and detection of labeled sacculi
E. coli cells (W3899) were incubated with photo-AeK-alkyne, collected and washed as
described above. Ice-cold cell suspension (1 ml) was slowly added to an equal volume of
63
boiling SDS solution (8% (w/v)), boiled for 30 min and left at room temperature overnight.
Sacculi were collected by ultracentrifugation (1 hr, 100 krpm, TLA 120.2 rotor, RT),
washed repeatedly with water, and resuspended in 200 µl 50 mM Tris-HCl pH 8.0. TCEP
(1 mM), CuSO4 (1 mM), TBTA (200 µM) and Alexa488-azide (50 µM) were added and
incubation was for 2 hrs at RT under gentle agitation. Sacculi were collected by
ultracentrifugation, washed with 50 mM Tris-HCl pH 8.0, 1% SDS and water, respectively,
resuspended in 50 µl of water and placed on microscope slides pre-coated with a thin flat
layer of 1% (w/v) low-melting-point agarose. Nikon Eclipse TE2000 confocal microscope
with a Pan Fluor 227 60.0x/1.40/0.21 oil objective was used for fluorescence analysis using
Argon laser excitation (488 nm) and Difference Interference Contrast (DIC) (Nomarsky
optics) was also used for detection.
In vivo crosslinking
E. coli cells (W3899) were grown and incubated with photo-AeK-alkyne as described
earlier. Cells were rapidly cooled in ice-water, collected by centrifugation at 4°C, washed
twice and resuspended in incubation buffer. For crosslinking, the samples were exposed
directly to the UV light at a wavelength of 302 nm by placing the lamp (8W UVP, Pierce)
above the samples at a distance of ~ 5 cm for 10 minutes at 4°C. Cells were subsequently
lysed by ultrasonication with a probe sonicator (Branson model 250, Branson Ultrasonics
Corp.) at 4°C. Sonication was done 5 times for 30 s each at 50 W. SDS was added to a final
concentration of 1% (w/v) and aliquots were stored at -80°C.
Detection of cross-linked proteins
In a typical experiment, 50 µg of cross-linked protein sample in 100 µl 50 mM Tris-HCl,
pH 8.0 containing 1% SDS (w/v), were reacted with either 50 µM Alexa488-azide or
biotin-azide in the presence of 1 mM TCEP, 1 mM CuSO4, 200 µM TBTA (final
concentrations). The reaction was for 2 hrs at RT. Five-fold concentrated SDS sample
buffer was added and proteins were separated by SDS-PAGE.
In gel detection of Alexa488 was performed on a Typhoon 9400 imaging system (GE
Healthcare Bio-Sciences AB, Uppsala, Sweden) using a 532 nm laser for excitation and a
526 nm band-pass filter for emission. Gels were subsequently stained with Coomassie
Brilliant Blue. For biotin detection, proteins were transferred to a nitrocellulose membrane
64
by Western blotting, blocked for 1 h with gelatin (3% (w/v), Biorad) in TBST (50 mM TrisHCl, 100 mM NaCl, 0,05% Tween 20, pH 8.0), incubated for 1 h with neutravidin- horse
radish peroxidase (HRP) (1 µg/mL; Pierce, Rockford, IL) in TBST containing gelatin (0.5%
(w/v)) and visualized, after washing, with ECL solutions.
Purification and identification of cross-linked proteins
Affinity purification - Protein solution containing biotinylated cross-link products
(approximately 50 µg protein, 100 µL) was incubated under gentle agitation with
neutravidin agarose beads (30 µL, Pierce) for 2 hrs at 4 °C. Beads were washed three times
with TBSS (50 mM Tris-HCl, 100 mM NaCl, 0,2% (w/v) SDS, pH 8.0) and bound proteins
were eluted by boiling for 10 min in 30 µL two-fold concentrated SDS-PAGE sample
buffer (125 mM Tris-HCl, 5% (w/v) SDS, 20% (v/v) glycerol, 100 mM dithiotreitol (DTT),
Bromophenol Blue, pH 6.8). Eluted proteins (10 µL) were separated by 12% SDS-PAGE
and the presence of the cross-link products was confirmed by Western blotting as described
above. For identification, 20 µL of the eluate was run over a distance of approximately 3
cm on a 12% SDS-PAGE gel. Following staining with Gelcode Blue (Pierce), bands were
excised and digested with trypsin essentially as described with minor modifications [14]. In
short, reduction and alkylation of protein disulfide bonds was conducted by incubating gel
pieces in 50 mM NH4HCO3 pH 8.5 containing 6.5 mM DTT and 55 mM iodoacetamide,
with intermediate washing and dehydrating steps using 50 mM NH4HCO3 pH 8.5 and
acetonitrile, respectively. Trypsin (150 ng) was added to the gel pieces in a final volume of
20 µL 50 mM NH4HCO3 pH 8.5 and digestion was carried out overnight at 37 ºC.
Following digestion, the supernatant was collected and peptides were additionally extracted
from the gel pieces using 5% (v/v) formic acid, after which both supernatants were pooled
for LC-MS/MS analysis.
Preparation of azide-derivatized beads and analysis of non-specific protein binding –
EAH-Sepharose 4B (GE Healthcare) and Magnabind-NH2 beads (Pierce) were converted
into azide form by the addition of the diazotransfer reagent imidazole-1-sulfonyl azide
hydrochloride [15]. Beads from 3 ml slurry, corresponding to ~30-36 µmol amine, were
collected, washed with water and resuspended in methanol (5 ml). Subsequently,
diazotransfer reagent (100 µmol) and 1 mol% CuSO4 were added and the reaction slurry
was stirred for 4 hrs at RT. The beads were washed twice with methanol and twice with 50
65
mM Tris-HCl pH 8.0. Conversion of amine groups was monitored by boiling the beads in
1% (w/v) ninhydrin in water for 5 min. Formation of azides was confirmed by conjugation
of TAMRA-alkyne to the beads in Tris-HCl pH 8.0 containing 1 mM TCEP, 1 mM CuSO4,
200 µM TBTA. After washing with methanol, beads were analyzed for fluorescence under
UV light. To test for non-specific binding, beads were collected from 10-100 µL slurry
(50%) and incubated with 50 µL E. coli protein lysate (~0,5 mg/ml) in 50 mM Tris-HCl,
100 mM NaCl, 1% (w/v) SDS, pH 8.0 for 3 hrs at RT with vigorous shaking. Beads were
washed three times with incubation buffer and boiled in two-fold concentrated SDS-sample
buffer for 10 min. The post-incubation supernatant and elution fractions were analyzed by
SDS-PAGE with Coomassie blue staining. The beads were stored at 4 ºC in Tris-HCl pH
8.0 containing 0.1% (w/v) NaN3.
Magnetic bead purification - Protein solution containing crosslink products (approximately
50 µg, 100 µL) was incubated with 30 µL Magnabind-N3 beads in the presence of 1 mM
TCEP, 1 mM CuSO4, 200 µM TBTA at RT for 2 hrs with shaking. Beads were collected
and washed once with 50 mM Tris-HCl, pH 7.5, 1 M NaCl, once with 50 mM Tris-HCl, pH
7.5, 8 M urea, 1% (w/v) SDS and twice with 50 mM NH4HCO3 pH 8.5. Reduction and
alkylation was carried out as described above, with two 50 mM NH4HCO3 pH 8.5 washing
steps in between. Proteins were digested at 37ºC overnight by the addition of 2 µg trypsin
in a final volume of 50 µL 50 mM NH4HCO3 pH 8.5. Supernatants were collected for LCMS/MS analysis.
LC-MS/MS analysis - Tryptic digests from gel pieces and bead suspensions were analyzed
by nano-scale LC-MS/MS by coupling an Agilent 1100 Series LC system to a LTQ XL
quadrupole ion trap mass spectrometer (Finnigan, San Jose, CA). Tandem mass spectra
were extracted and charge state deconvoluted by BioWorks (Thermo Scientific, Waltham,
MA; version 3.3). All MS/MS samples were analyzed using Mascot (Matrix Science,
London, UK; version 2.2.03) set up to search the SwissProt_53.2 E. coli database (8976
entries) with a parent ion tolerance of 0.5 Da and a fragment ion tolerance of 0.9 Da. The
iodoacetamide derivative of cysteine and oxidation of methionine were specified as fixed
and variable modification, respectively. Scaffold (version 01_07_00, Proteome Software,
Portland, OR) was used to validate MS/MS based peptide and protein identifications.
Peptide identifications were accepted if they could be established at greater than 95.0%
66
probability as specified by the Protein Prophet algorithm [16]. Proteins that contained
similar peptides and could not be differentiated based on MS/MS analysis alone were
grouped to satisfy the principles of parsimony.
Results
We had shown in Chapter 2 that fluorescently labeled murein tripeptide was taken up by E.
coli and incorporated into the cell wall via the peptidoglycan recycling pathway. In order to
comprehensively identify proteins along the metabolic pathway of peptidoglycan in vivo, a
murein tripeptide analogue was synthesized containing a photoreactive moiety and an
alkyne functionality: photo-AeK-alkyne (Fig. 2). The strategy was to crosslink enzymes
interacting with the cell wall and/or its intermediates in vivo and to utilize the alkyne
functionality for detection and purification of the crosslink products through in vitro
conjugation of azide-modified reporter tags (Fig. 3). To confirm that photo-AeK-alkyne is
also metabolically utilized, we tested for the formation of labeled Lipid II and labeled
peptidoglycan. E. coli wild-type cells (W3899) were incubated with the photoactivatable
tripeptide and Lipid II was specifically extracted by acidic Bligh and Dyer, after which the
Figure 2. Structure of photo-AeK-alkyne (TFA•H2N-Ala-γ-D-Glu-L-Lys(ε−N-p-azido-L-Phe-N-5hexynamide)-COOH).
67
Cu(I)
UV
R
R
Photocrosslinking group
R
Alexa488, biotin, magnetic beads
Figure 3. Proteins interacting with peptidoglycan (subunits) are covalently bound following UVphotocrosslinking. Reporter tags or beads are subsequently attached via click chemistry for detection
and enrichment purposes.
extract was subjected to click chemistry using Alexa488-azide. TLC analysis with UV and
iodine detection showed that the extracted Lipid II was fluorescently labeled and had a
higher Rf value than native Lipid II, which is indicative for the modification at the lysine
residue (Fig. 4). Correspondingly, following incubation with photo-AeK-alkyne, sacculi
Figure 4. TLC analysis of Lipid II extraction, performed as described in Materials and methods, after
incubation of wild-type E. coli (W3899) with photo-AeK-click. The extract was subjected to click
chemistry using Alexa488-azide to confirm the presence of alkyne derivatized Lipid II. Lipid II
standard (left lanes) and the presence of Alexa488-labeled Lipid II in the extract (right lanes) are
visualized by iodine staining and UV light, respectively. As a reference, in vitro synthesized AlexaLipid II is shown.
68
were isolated and subsequently treated with Alexa488-azide under click chemistry
conditions. Confocal microscopy showed that sacculi were fluorescently labeled, whereas
no fluorescence was observed in sacculi that were incubated with the fluorescent tag after
they were isolated from untreated cells (Fig 5). This shows that the alkyne-labeled
tripeptide analogue is also metabolically incorporated into peptidoglycan of E. coli. Next,
the in vivo crosslinking experiment was commenced by allowing incorporation of photoAeK-alkyne, followed by washing of the cells to get rid of peptide not taken up by the cells.
Cells were exposed to UV light for 10 minutes on ice and subsequently lysed by
ultrasonication. Click chemistry was performed to detect the cross-linked proteins using
Alexa488-azide as fluorescent reporter tag. Proteins were resolved by SDS-PAGE and ingel fluorescence scanning showed multiple fluorescently labeled proteins in the UVexposed sample (Fig. 6, bottom). The most intense of these bands were also visible when
UV illumination had been absent, most probably caused by background light. The
appearance of these protein bands can be attributed to the incorporation of photo-AeKalkyne, since in the absence of crosslinker negligible background signal was observed (Fig.
6, bottom). The pattern and intensities of the bands showed significant differences from
those observed after Coomassie staining (Fig. 6, top) demonstrating the specificity of the
crosslinking approach.
Figure 5. Fluorescent (left) and DIC (right) image of E. coli sacculus. E. coli wild-type cells (W3899)
were incubated with photo-AeK-alkyne and peptidoglycan was isolated by boiling cells in SDS.
Sacculi were subjected to click chemistry using Alexa488-azide.
69
Figure 6. In vivo cross-linking of proteins by (metabolized forms of) photo-AeK-alkyne. SDS-PAGE
of protein lysates obtained from wild-type cells (W3899) incubated with photo-AeK-alkyne and either
exposed to UV light or left in the dark. Proteins from untreated cells were used as a control. All
samples were subsequently incubated with Alexa488-azide in the presence of Cu(I). Fluorescence
scan showing selective detection of in vivo cross-linked proteins ligated to Alexa488-azide (bottom).
Gels were subsequently stained with Coomassie blue (top).
To identify the crosslinked proteins, they were initially reacted with a biotin-azide probe.
This allowed for affinity purification on neutravidin–agarose resin. First, biotinylation of
crosslink products was confirmed by detection with neutravidin-HRP on Western blot (Fig.
7). The pattern of the bands resembled that observed for the fluorescently labeled proteins,
although detection by fluorescence showed a higher resolution and dynamic range. Again,
70
some biotinylated proteins were observed when UV-illumination had been omitted, while
negligible signal was observed in the absence of crosslinker, except for one band. This band
at approximately 21 kDa corresponds to biotin carboxyl carrier protein (BCCP), the only
endogenous biotinylated protein in E. coli. Following the biotinylation step, proteins were
incubated with neutravidin beads. Several washing steps were then conducted to remove
proteins bound non-specifically to the beads. The proteins bound to the beads were eluted
by boiling in two-fold concentrated SDS sample buffer and separated by SDS-PAGE. The
capture of biotinylated proteins by the neutravidin beads was highly efficient as was
Figure 7. Western blot analysis of crosslink products in E. coli protein lysates. Sample preparation
was as described in Figure 3, except biotin-azide was used as reporter probe. Following SDS-PAGE,
proteins were blotted to nitrocellulose membranes. Biotinylated proteins were detected using
Neutravidin-HRP and enhanced chemiluminescence.
assessed by Western blot (data not shown). Following in-gel trypsin digestion, proteins
were identified by LC-MS/MS (Table 1). Strikingly, the control sample not incubated with
the photoactivatable tripeptide showed a significant number of proteins that apparently had
bound non-specifically to the neutravidin beads. Moreover, the fraction of proteins that
were not present in the control and therefore most likely specifically purified, was obscured
by the contaminating non-specifically purified proteins. In an attempt to circumvent this
71
Table 1. LC-MS/MS analysis of crosslink products, biotinylated using click chemistry and purified on
neutravidin beads. Proteins that were also identified in the negative control are not shown.
gi number
15800320
1827931
146079
26245929
442835
15804231
15802330
15804731
37953960
15803175
26248017
26249321
26247975
15799906
13591815
9632520
26249616
26249305
15830633
15722333
26249651
606350
1305720
2226075
15804653
15802395
3929340
10955324
136136
15800432
228463
15802464
15802375
26250542
Protein
Alkyl hydroperoxide reductase
Acetyl-CoA carboxylase
Glyceraldehyde-3-phosphate dehydrogenase
Transaldolase
Maltodextrin-binding
Hypothetical protein Z5121
Putative replication protein
Aspartate ammonia-lyase
Proline permease
Putative enzyme
Putative TPR repeat protein
Hypothetical protein c3486
6-Phosphofructokinase isozyme
Hypothetical protein z0257
Extended spectrum β-lactamase
Hypothetical 933wp54
Arylsulfate sulfotransferase
Hypothetical protein c3470
Hypothetical protein
Trwk
Hypothetical protein c3820
Gluconate permease
Prs-associated putative membrane protein
Trbc
Hypothetical protein z5660
DNA cytosine methylase
Colicin-10
Toxin B
Transposase
Succinyl-CoA synthetase
AcrF
Shikimate transport protein
Flagellar assembly protein
Uroporphyrin-III synthetase
72
MW (Da)
20862
8743
33183
37676
40682
14047
107628
54714
32038
38161
75189
5633
32831
19582
40773
44991
66700
55673
8061
93943
42689
46107
47413
23433
60970
53773
53310
362921
28442
41652
104851
48182
25867
28088
Mascot
ID score
125
89
75
66
51
49
48
45
45
40
39
39
38
38
38
37
37
37
37
36
36
36
36
36
36
36
35
35
35
35
34
34
34
33
120410
89109564
16131552
15802965
40846
9507812
56123290
57165055
537084
26250929
29420393
46451705
26250979
16129697
9632520
8569539
117552
15802956
51247112
15804552
56266653
227912
15834215
37528724
47242
15803040
15804423
15803723
15802428
25815152
290505
16129520
37927527
15804781
42043
682767
11514100
26247302
Fimbrial protein (P adhesin)
Hypothetical protein (putative
membrane/lipoprotein)
DNA-binding transcr. Regulator
Predicted acyltransferase
Acetohydroxy acid synthase
TraG
NnaC
o121 (putative ubiquinone biosynthesis protein
UbiB)
Mg2+ transport protein Mgt
Hypothetical protein yjdB
Wbqd
MobA
Hypothetical protein c5171
Envelope stress induced protein
Flagellin
Cyanase
Chaperone cs3-1
Thiosulfate binding protein
Alcohol dehydrogenase
Phosphoenolpyruvate carboxylase
DNA pacase
S fimbrial adhesin (FocD)
Putative transcription regulator
Putative glycosyltransferase
Alkaline phosphatase
Predicted enzyme
Uridine phosphorylase
GTPase
Unknown protein
Hypothetical protein
Hypothetical protein (AsmA family)
Qin prophage
MobC
L-Ascorbate 6-phosphate lactonase
MukB
MccD
MutS
Lysis protein
73
32664
33
10512
27368
20587
17938
102536
48539
33
33
33
33
32
32
14090
99861
64281
16318
55405
10019
18187
47555
17015
27062
37591
44663
99498
45278
95779
16206
38580
47242
43514
27304
43487
37674
25843
62258
9400
29152
40434
178138
31601
89303
12270
32
32
32
32
32
32
32
32
32
32
32
31
31
31
31
31
31
31
31
31
31
30
30
30
30
30
30
30
30
30
30
26246063
537218
16131450
10955493
CueO (SufI)
Predicted DNAse
Predicted transporter
Site-specific recombinase
58394
23706
35948
43153
30
30
30
30
problem, we developed an alternative purification approach in which cross-linked proteins
were purified by directly binding them covalently onto beads through click chemistry. For
this, sepharose and Magnabind beads derivatized with amine groups were incubated with
the diazotransfer reagent imidazole-1-sulfonyl azide hydrochloride to convert the amines
into azides. Conversion was confirmed by the lack of staining of the beads with ninhydrin
and by monitoring fluorescence of the beads following click reaction with TAMRA-alkyne
conjugate (data not shown). Beads were subsequently tested for non-specific binding by
incubating with E. coli cell lysate followed by washing and elution by boiling in SDS
sample buffer. SDS-PAGE showed that for the amine-derivatized beads and the sepharose-
Figure 8. SDS-PAGE analysis of non-specific protein binding to beads. Magnetic azide beads from
25, 50 and 100 µL of slurry (50%) were incubated with 50 µL E. coli protein lysate (~0,5 mg/ml).
Protein quantity in post-incubation supernatant was unchanged with increasing number of beads (s25,
s50, s100) and no protein bands were visible after extensive boiling in SDS sample buffer (e25, e50,
e100).
74
azide beads, increasing the amount of beads resulted in significant protein loss in the
supernatant and visible protein bands in the elution fraction, indicating that both the amine
functionality and the sepharose matrix are a source of non-specific protein binding. In
contrast, magnetic azide beads were shown to be susceptible to very little non-specific
protein binding (Fig. 8) and were used for subsequent purification of cross-linked proteins.
This was achieved by incubating the proteins with the Magnabind azide beads in the
presence of copper (I). The beads were washed and bound proteins were subjected to onbead trypsin digestion and subsequent identification by peptide mass fingerprinting. LCMS/MS data showed a vastly increased number of identified proteins, each with a much
higher identification score, and relatively few proteins were present in the control (Table 2).
Table 2. LC-MS/MS analysis of crosslink products purified via click chemistry on azide-derivatized
magnetic beads. Mascot scores of proteins that were identified in the negative control are also listed.
acc. Number
Swiss-prot
LPP_ECOLI
OMPA_ECOLI
OMPC_ECOLI
RPOC_ECOLI
OMPC_ECOL6
CH60_ECO24
ODP1_ECO57
Protein
Major outer membrane lipoprotein
Outer membrane protein A
Outer membrane protein C
DNA-directed RNA polymerase subunit beta'
Outer membrane protein C
60 kDa chaperonin
Pyruvate dehydrogenase E1 component
Dihydrolipoyllysine-residue acetyltransferase
ODP2_ECOLI component of pyruvate dehydrogenase complex
ATMA_ECOLI Magnesium-transporting ATPase, P-type 1
TNAA_ECOHS Tryptophanase
RS2_ECO57
30S ribosomal protein S2
RS11_ECO24 30S ribosomal protein S11
PAL_ECOLI Peptidoglycan-associated lipoprotein
RL6_ECO24 50S ribosomal protein L6
KPRS_ECOLI Ribose-phosphate pyrophosphokinase
DLDH_ECOLI Dihydrolipoyl dehydrogenase
OMPF_ECOLI Outer membrane protein F
PTTBC_ECOLI PTS system trehalose-specific EIIBC component
RL2_ECO24 50S ribosomal protein L2
RL10_ECO24 50S ribosomal protein L10
EFTU_ECOLI Elongation factor Tu
75
MW
(Da)
8375
37292
40343
155918
41200
57464
99948
score control
650 259
281 61
265
250
242
240
198
66112
99973
53197
26784
13950
18869
18949
34425
50942
39309
51389
29956
17757
43457
190
172
163
140
132
130
129
126
113
113
108
106
102
100
117
147
93
ATPA_ECOHS
NUPC_ECOLI
RS7_ECO24
RPOB_ECOHS
PLSB_ECOHS
RS9_ECO24
PHSM_ECOLI
RS18_ECO24
RL20_ECO24
ODO1_ECO57
G3P1_ECO57
DHNA_ECOLI
ATP synthase subunit alpha
Nucleoside permease nupC
30S ribosomal protein S7
DNA-directed RNA polymerase subunit beta
Glycerol-3-phosphate acyltransferase
30S ribosomal protein S9
Maltodextrin phosphorylase
30S ribosomal protein S18
50S ribosomal protein L20
2-oxoglutarate dehydrogenase E1 component
Glyceraldehyde-3-phosphate dehydrogenase A
NADH dehydrogenase
55416
43504
17593
150937
91609
14847
90865
9038
13489
105566
35681
47557
99
96
96
95
91
89
87
87
86
83
83
78
123
74
Discussion
In this study we used a photoactivatable analogue of murein tripeptide, photo-AeK-alkyne,
to crosslink enzymes interacting with peptidoglycan and its precursors in vivo. As was
shown in chapter 2, NBD-labeled murein tripeptide entered E. coli cells and was
metabolically incorporated into the cell wall. Photo-AeK-alkyne, containing a somewhat
larger label, was also incorporated via the same metabolic pathway as evidenced by the
formation of derivatized Lipid II and sacculi. The crosslinking of proteins in vivo offers the
big advantage to study the interactions in their native cellular environment and an
additional benefit of our method is that it allows the crosslinking of proteins along virtually
the complete metabolic pathway of peptidoglycan. The presence of an alkyne group
allowed for the detection and purification of the crosslinked proteins via click chemistry.
Biotinylation followed by capture using immobilized biotin-binding proteins, such as
neutravidin, is a commonly used method for the purification of biomolecules. It is highly
sensitive but also prone to high background and false positives due to non-specific binding.
LC-MS/MS analysis of the purified biotinylated crosslink products showed that each
sample, irrespective of treatment with the labeled tripeptide and/or UV activation,
contained the same subset of proteins, many of which were identified with the highest
scores. Click chemistry is known to be highly selective, therefore most likely ruling out the
possibility of non-specific protein labeling by the biotin-azide probe. This was confirmed
by Western blot analysis with neutravidin-HRP/ECL detection, which showed negligible
76
signal in the negative control. However, SDS-PAGE analysis with Coomassie staining of
the purified crosslink products showed the same pattern of bands in each sample (data not
shown), and unlike the patterns observed on Western blot. This demonstrates the presence
of a relatively high amount of non-specifically bound proteins to the neutravidin beads that
are apparently only detached from the beads during the harsh elution conditions. Although
they can easily be distinguished, the main disadvantage is that their relatively high
abundance is likely to obscure specifically crosslinked proteins that are present in much
lower quantities. Non-specific contamination could possibly be reduced by mild elution
conditions using biotin containing buffers, however, this will likely also decrease recovery
of the crosslink products. Despite the drawbacks associated with the biotin purification,
various interesting crosslinked proteins were identified (Table 1). AcrF is part of the
multidrug efflux system AcrEF-TolC, which has also been implicated to play a role during
cell division [17]. Interestingly, overexpression of AcrEF was shown to suppress division
defects caused by mutations in envC, a gene encoding a septal murein hydrolase [18, 19].
How this is achieved remains unclear. TraG and TrwK are membrane proteins involved in
type IV secretion with TraG having a possible peptidoglycanase domain [20]. An
uncharacterized protein from the yjdB gene was identified. The protein encoded from this
gene, EptA contains a hydrolase domain and is predicted to be involved in Z-ring formation
during cell division. Another protein involved in cell division is SufI (FtsP). Its exact
function is unclear, but it appears to be required for growth under stress conditions and is
capable of suppressing growth defects in several division mutants. It has been suggested
that it stabilizes divisome assembly, possibly through interaction with peptidoglycan [21].
Several hypothetical proteins were found, including a putative glycosyltransferase, a
conserved protein highly homologous to a putative lipoprotein of Burkholderia species, and
a protein of ~46 kDa that has some homology to a putative D-alanyl-D-alanine
carboxypeptidase of C. luteolum.
To directly utilize the high specificity of the azide-alkyne cycloaddition for purification
purposes, azide-derivatized beads were tested. The direct capture of crosslink products onto
beads via click chemistry eliminates the need for an additional high affinity binding step,
such as biotin-neutravidin, limiting the chance for non-specific protein binding. Azide
beads could be prepared from the amine-form in a straight-forward approach using the
diazotransfer reagent imidazole-1-sulfonyl azide hydrochloride [15]. Magnetic beads were
shown to intrinsically have low protein binding capacity. The covalent attachment of the
77
proteins via click chemistry also allowed for harsh washing conditions reducing
contamination, as was observed following on-bead digestion and MS identification. Taken
together, the azide functionalized magnetic beads could serve as a powerful purification
tool for alkyne-modified proteins.
Identification of crosslink products purified with azide-beads, listed in Table 2, resulted in
some expected proteins, such as the major lipoproteins Lpp and Pal, several major outer
membrane proteins (OmpA, OmpC, OmpF). Many ribosomal proteins were identified as
well as other high abundant cytoplasmic proteins. Some proteins were still present in the
negative control, including Lpp and OmpA, but their scores were considerably higher in the
sample subjected to labeling, indicating that they were crosslinked by labeled
peptidoglycan units. In contrast, the other proteins also present in the negative control, such
as elongation factor Tu and NADH dehydrogenase had similar or lower scores and were
thus non-specific. As our data show, particularly cytoplasmic proteins present in the cell in
high abundance were cross-linked, indicating that the free peptide present in the cytoplasm
is liable to be in close proximity of these proteins resulting in non-specific crosslink
products. The danger therein lies in that these frequently occuring non-specific cross-links
could obscure the specific interactions that are evidently captured in lower numbers. It
would be a daunting task to try to attain an optimal balance between high efficiency and
specificity of cross-linking by the labeled cell wall units/precursors. Since the cytoplasmic
steps of peptidoglycan biosynthesis have been relatively well characterized, the
identification of a higher number of specific cross-link products could possibly be achieved
by fractionating the cell into membrane and cell wall components. Preliminary experiments
using this strategy showed promising data, however, reliability of protein identifications
was hampered most likely due to low quantity of material obtained for mass spectrometry
analysis. As with most proteomic studies, the challenge lies in data interpretation and
possible interactions should be validated by orthogonal methods. In conclusion, the here
reported results show that our in vivo crosslinking approach using photo-activatable cell
wall precursors offers great potential as a tool for the elucidation of proteins interacting
with peptidoglycan or its precursors/breakdown products. This will further deepen our
understanding on how interacting protein machineries coordinate cell wall metabolism with
major cellular processes such as cell division and among the newly identified proteins there
could be promising new targets for antibiotic action.
78
Acknowledgments
Martina O’Flaherty and Mirjam Damen are acknowledged for mass spectrometry analysis.
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
Errington, J., R.A. Daniel, and D.J. Scheffers, Cytokinesis in bacteria. Microbiol
Mol Biol Rev, 2003. 67(1): p. 52-65.
Nanninga, N., Cell division and peptidoglycan assembly in Escherichia coli. Mol
Microbiol, 1991. 5(4): p. 791-5.
Boneca, I.G., The role of peptidoglycan in pathogenesis. Curr Opin Microbiol,
2005. 8(1): p. 46-53.
Zahrl, D., et al., Peptidoglycan degradation by specialized lytic transglycosylases
associated with type III and type IV secretion systems. Microbiology, 2005. 151(Pt
11): p. 3455-67.
van Alphen, W., N. van Seim, and B. Lugtenberg, Pores in the outer membrane of
Escherichia coli K12: involvement of proteins b and e in the functioning of pores
for nucleotides. Mol Gen Genet, 1978. 159(1): p. 75-83.
Chin, J.W., et al., Addition of a photocrosslinking amino acid to the genetic code
of Escherichiacoli. Proc Natl Acad Sci U S A, 2002. 99(17): p. 11020-4.
Tanaka, Y. and J.J. Kohler, Photoactivatable crosslinking sugars for capturing
glycoprotein interactions. J Am Chem Soc, 2008. 130(11): p. 3278-9.
Speers, A.E., G.C. Adam, and B.F. Cravatt, Activity-based protein profiling in vivo
using a copper(i)-catalyzed azide-alkyne [3 + 2] cycloaddition. J Am Chem Soc,
2003. 125(16): p. 4686-7.
Saghatelian, A., et al., Activity-based probes for the proteomic profiling of
metalloproteases. Proc Natl Acad Sci U S A, 2004. 101(27): p. 10000-5.
Gubbens, J., et al., Photocrosslinking and click chemistry enable the specific
detection of proteins interacting with phospholipids at the membrane interface.
Chem Biol, 2009. 16(1): p. 3-14.
MacKinnon, A.L., et al., Photo-leucine incorporation reveals the target of a
cyclodepsipeptide inhibitor of cotranslational translocation. J Am Chem Soc,
2007. 129(47): p. 14560-1.
Olrichs, N.K., et al., A novel in vivo fluorescent labeling approach sheds new light
on peptidoglycan synthesis in Escherichia coli. submitted.
Guan, Z., S.D. Breazeale, and C.R. Raetz, Extraction and identification by mass
spectrometry of undecaprenyl diphosphate-MurNAc-pentapeptide-GlcNAc from
Escherichia coli. Anal Biochem, 2005. 345(2): p. 336-9.
Wilm, M., et al., Femtomole sequencing of proteins from polyacrylamide gels by
nano-electrospray mass spectrometry. Nature, 1996. 379(6564): p. 466-9.
Goddard-Borger, E.D. and R.V. Stick, An efficient, inexpensive, and shelf-stable
diazotransfer reagent: imidazole-1-sulfonyl azide hydrochloride. Org Lett, 2007.
9(19): p. 3797-800.
Nesvizhskii, A.I., et al., A statistical model for identifying proteins by tandem
mass spectrometry. Anal Chem, 2003. 75(17): p. 4646-58.
79
17.
18.
19.
20.
21.
Lau, S.Y. and H.I. Zgurskaya, Cell division defects in Escherichia coli deficient in
the multidrug efflux transporter AcrEF-TolC. J Bacteriol, 2005. 187(22): p. 781525.
Klein, J.R., B. Henrich, and R. Plapp, Molecular analysis and nucleotide sequence
of the envCD operon of Escherichia coli. Mol Gen Genet, 1991. 230(1-2): p. 23040.
Hara, H., et al., Identification and characterization of the Escherichia coli envC
gene encoding a periplasmic coiled-coil protein with putative peptidase activity.
FEMS Microbiol Lett, 2002. 212(2): p. 229-36.
Kohler, P.L., et al., AtlA functions as a peptidoglycan lytic transglycosylase in the
Neisseria gonorrhoeae type IV secretion system. J Bacteriol, 2007. 189(15): p.
5421-8.
Samaluru, H., L. SaiSree, and M. Reddy, Role of SufI (FtsP) in cell division of
Escherichia coli: evidence for its involvement in stabilizing the assembly of the
divisome. J Bacteriol, 2007. 189(22): p. 8044-52.
80
Chapter 4
Probing the Lipid II binding site of
Escherichia coli penicillin-binding
protein 1b.
Abstract
The peptidoglycan cell wall is a giant molecule surrounding bacterial cells consisting of
alternating N-acetylglucosamine (GlcNAc) and N-acetylmuramic acid (MurNAc) units,
interlinked via peptide moieties that are attached to the MurNAc residue. The cell wall is
essential for cell survival and its biosynthesis machinery provides many targets for
antibacterial agents. During the extracellular stage of peptidoglycan biosynthesis,
transglycosylases catalyze the formation of linear glycan chains using the membrane-bound
precursor Lipid II as substrate. The majority of transglycosylation activity is performed by
bifunctional penicillin binding proteins (PBPs), which also catalyze the final
transpeptidation step. Transglycosylases contain two substrate binding sites: one for the
lipid-linked growing glycan chain, which acts as a glycosyl donor and one for the glycosyl
acceptor Lipid II. The only known direct transglycosylase enzyme inhibitor, moenomycin,
achieves its inhibitory action by binding to the donor site. For the development of novel
transglycosylase targeting compounds, it is important to fully characterize the Lipid II
binding site. We used in vitro photo-crosslinking in combination with mass spectrometry
techniques to study the interaction between Lipid II and the bifunctional transglycosylase
PBP1b of Escherichia coli. Radiolabeled Lipid II derivatized with a photo-activatable
crosslinker was prepared and utilized as a substrate by PBP1b in vitro. When
polymerization was inhibited by moenomycin, Lipid II was shown to crosslink to one major
site in the protein. Vancomycin-sepharose was used to affinity-purify crosslinked peptides
from the tryptic digest. Mass spectrometry analysis suggested that Lipid II was covalently
bound to a section of the transglycosylase domain containing highly conserved residues,
which possibly encompasses the specific binding site of Lipid II.
Introduction
The peptidoglycan cell wall is essential for bacterial cell survival and its biosynthesis
machinery is an ideal target for antibiotics. The final stages of peptidoglycan synthesis,
which involves incorporation of the membrane bound precursor, Lipid II into the preexisting peptidoglycan, is catalyzed by the penicillin-binding proteins (PBPs) [1]. Class A
(bifunctional) PBPs possess both enzymatic activities required for peptidoglycan assembly
from the Lipid II substrate; the polymerization of disaccharide units of Lipid II into
82
elongated glycan chains (transglycosylase activity) and the crosslinking of the peptide side
chains from adjacent glycan strands (transpeptidase activity). Escherichia coli has three
class A PBPs; 1a, 1b and 1c. PBP1b is encoded by the ponB gene and it exists in three
molecular forms (α, β, γ), which vary slightly in the length of their N-terminal cytoplasmic
part [2, 3]. The transglycosylase domain consists of amino acids 198-435 and the
transpeptidase domain of amino acids 447-780. Both domains are located in the periplasm,
with the transglycosylation domain anchored to the plasma membrane via a transmembrane
helix with a short N-terminal cytoplasmic tail. PBP1b is the major enzyme for
peptidoglycan synthesis in E. coli [4, 5] and is known to form dimers in vivo [6-8]. It has
been shown to polymerize Lipid II in vitro, with higher activity in conditions favoring
dimerization [9].
Penicillin-binding proteins in particular are excellent targets for antibiotics because they are
essential, no equivalent enzymes are present in mammalian cells and they are highly
accessible due to their extracellular location. PBPs are the target of the β-lactam antibiotics,
which covalently attach to transpeptidase domains, inhibiting the formation of peptide
cross-bridges and thereby incorporation of nascent peptidoglycan into the existing cell wall
[1]. The emerging resistance to these antibiotics has called for more in depth knowledge on
the PBPs, with a renewed focus on the transglycosylase activity. The only well-studied
direct transglycosylase enzyme inhibitor is moenomycin. It has very potent antibacterial
activity against Gram-positive bacteria but is not effective in humans, due to poor
pharmacokinetic properties. This leaves much room for the discovery of new
transglycosylase inhibiting compounds, which could be less prone to resistance, as the
aminosugar residues of peptidoglycan show very little structural variation in contrast to the
peptide part.
In recent years, considerable progress has been made in understanding the
transglycosylation mechanism. The lipid linked growing polymer was found to function as
glycosyl donor with Lipid II acting as the glycosyl acceptor (Chapter 1, Fig. 4) [10, 11].
Furthermore, a number of residues are highly conserved among transglycosylases from
different bacterial species, and several of these were shown to be important for catalytic
activity [12]. The long-awaited elucidations of transglycosylase crystal structures, including
some in complex with moenomycin, revealed that the transglycosylase domain contains an
extended binding site cleft between a globular head subdomain and a flexible jaw region
[13, 14]. Moenomycin structurally mimics the lipid-linked tetrasaccharide part (Lipid IV)
83
of the growing chain and prevents transglycosylation by occupying the donor binding site.
The interaction between enzyme and inhibitor has been characterized in great detail and this
structural information is now widely used in screens for and the design of novel inhibitors
[15, 16]. Much less is known about the transglycosylase interaction with Lipid II, as this
substrate is difficult to obtain this molecule in workable quantities and handling in
enzymatic assays is complicated by its hydrophobic character. Nevertheless, it is essential
to fully elucidate this interaction as it will provide a broad-range target for transglycosylase
inhibition. In this study, we have used an in vitro photo-crosslinking approach to get insight
into the binding site of Lipid II in the transglycosylase domain of Escherichia coli PBP1b.
It is shown that when radioactive Lipid II derivatized with a photo-crosslinking group is
incubated with PBP1b in the presence of moenomycin, it is covalently captured to one
major site upon exposure to UV light. Affinity-purification of tryptic peptides crosslinked
to Lipid II was achieved using vancomycin-sepharose. Analysis by MALDI-TOF mass
spectrometry indicated that Lipid II was crosslinked to the tryptic peptide with the amino
acid sequence His236-Arg260, which is located in the transglycosylase domain and
contains conserved residues essential for catalytic activity, and could therefore encompass
the Lipid II binding site.
Materials and Methods
Materials
Escherichia coli PBP1b (γ-form), His-tagged at an N-terminal extension of six amino acids,
H6(GSHMASM46-N844), that was purified as described, was a kind gift from Prof. W.
Vollmer, Newcastle University. The enzyme was stored at a concentration of 1.34 mg/ml in
a buffer containing 20 mM sodium phosphate, pH 6, 300 mM NaCl, 10 mM MgCl2, 10%
(w/v) glycerol, 1% (w/v) Triton X-100, 0.02% (w/v) NaN3 at -20° C.
Preparation of [14C]Lipid II-ASA
Isolation and labeling of UDP-MurNAc-pentapeptide
UDP-MurNAc pentapeptide (lysine form) was isolated from Staphylococcus simulans and
purified as decribed [17] and labeled at the lysine residue using of N-Hydroxysuccinimidyl4-azidosalicylic acid (NHS-ASA). 70 mg UDP-MurNAc-pentapeptide was dissolved in 3.2
ml acetonitrile/100 mM NaHCO3, pH 9 (7:3 v/v). 20 mg NHS-ASA was dissolved in DMF
84
and slowly added under constant stirring to the UDP-MurNAc-pentapeptide solution. All
experiments involving the photoactivatable crosslinker were performed under subdued
light. Incubation was for 3 hours at room temperature and the reaction product was purified
by C18 reversed-phase HPLC using a linear gradient from 50 mM ammonium bicarbonate
to 100% methanol. Identity of UDP-MurNAc-pentapeptide-ASA was confirmed by
MALDI-TOF mass spectrometry and its concentration was determined by phosphorus
analysis [18].
Synthesis and purification of [14C]Lipid II-ASA
Lipid I-ASA was synthesized by incubating M. flavus vesicles (~1,5 µmol lipid-Pi) with 2,5
µmol UDP-MurNAc-pentapeptide-ASA, 600 nmol undecaprenylphosphate, in 4.5 ml of
buffer containing 100 mM Tris-HCl, pH 8, 5 mM MgCl2, and 1% (w/v) Triton X-100. The
suspension was incubated at room temperature for 2 h, followed by extraction of the lipids
by 6 ml of butanol/6 m pyridine-acetate, pH 4.2. The butanol (top) phase was collected
after brief centrifugation and washed with 4 ml of water. Reaction products were analyzed
by TLC using chloroform/methanol/water/ammonia (88:48:10:1), and the spots were
visualized by iodine vapor.
Purification of Lipid I-ASA was performed using a DEAE-cellulose column (acetate form)
of 4 × 2.5 cm (height × diameter)[19]. Lipid I-ASA was eluted at 500 mM ammonium
bicarbonate if a linear gradient was used from chloroform/methanol/water (2:3:1) to
chloroform/methanol/700 mM ammonium bicarbonate (2:3:1).
Lipid I-ASA was converted into [14C]Lipid II-ASA by MurG, which was purified as
described previously [20]. First, Lipid I-ASA (100 nmol) was dissolved in 200 µl of a
buffer containing 100 mM Tris-HCl, pH 8, 1 mM MgCl2, 1% (w/v) Triton X-100.
Subsequently, 25 µCi of [14C]UDP-GlcNAc (10.5 GBq/mmol, Amersham Biosciences) and
10 µg of MurG were added. The mixture was incubated for 3 hours at room temperature
and Lipid II-ASA was extracted with 300 µl of butanol, 6 M pyridine acetate, pH 4.2, and
purified as described for Lipid I-ASA with elution of Lipid II-ASA at 600 mM ammonium
bicarbonate. [14C]Lipid II-ASA had a specific activity of 605 dpm/µmol and was stored in
chloroform/methanol 1:1 (v/v) at a concentration of 500 µM at -20 °C until use.
85
Transglycosylation and photo-crosslinking
[14C]Lipid II-ASA (20 nmol) was vacuum-dried and dissolved on ice in 5 µl of 1% Triton
X-100 for 10 min. The reactions were performed in buffer containing 10 mM
HEPES/NaOH, pH 7.5, 3 mM MgCl2, 3.4% glycerol, 0.18% Triton X-100, and 150 mM
NaCl. Optionally, moenomycin (50 µM) was present to inhibit transglycosylation. PBP1b
was present at a concentration of 5 µM, and the total reaction volume was 500 µl. The
reaction mixture was incubated at 37 °C for 1 hour. Polymerization of [14C]Lipid II-ASA
was monitored by thin-layer chromatography (TLC) essentially as described [21]. Briefly,
the reaction mixture was spotted onto a Cellulose F TLC plate (Merck) and developed with
isobutyric acid/1 M ammonia (5:3 (v/v)) at 4 °C. The TLC plate was air-dried before
exposure overnight to a storage-phosphor screen (Kodak) and radioactivity was detected
using the BioRad Personal Molecular Imager.
Crosslinking was induced by UV irradiation at 305 nm for 10 minutes. Optionally,
polymerized glycan strands were digested for the removal of non-crosslinked residues by
the addition of 10 µg hen egg-white lysozyme, followed by incubation at 37 °C for 3 hours.
SDS-PAGE analysis was performed using 12.5 % (w/v) acrylamide. For autoradiography,
gels were fixed for 20 minutes in 50 % (v/v) aqueous methanol containing 10% (v/v) acetic
acid, and subsequently washed with water for 45 minutes. Gels were soaked in Amplify®
enhancer solution for 20 minutes and dried at 60 °C for 2.5 hours and exposed for
approximately 2 days to Kodak BioMax MS films at -80°C.
For peptide analysis, the pH was adjusted to 4.8 by the addition of HCl and the sample was
boiled for 10 minutes to remove the polyprenyl tail of Lipid II. After cooling for 5 min,
protein was separated from non-crosslinked Lipid II headgroups by filtration through a 30
kDa MWCO filter (Millipore). The supernatant was diluted to a volume of 500 µl by the
addition of 50 mM ammonium bicarbonate, pH 7.8 before 10 µg ultrapure trypsin
(Promega) was added (~1:20 (w/w) trypsin:PBP1b ratio). Proteolytic digestion was allowed
to proceed overnight at 37 °C. The protein digest containing crosslinked peptides were
analyzed by reversed phase HPLC on a C18 column (5 µm particle size, 250 x 4.6 mm
inner diameter) using a linear gradient from water containing 0.1% (v/v) TFA to 100%
methanol with UV (214 nm) and flow-through radioactivity detection (Packard 500TR
FSA).
86
Preparation of vancomycin-sepharose and subsequent affinity purification.
Activated CH sepharose 4B (1 g) was suspended in 50 ml 1 mM HCl and washed four
times on a glass-sintered filter with 50 ml 1 mM HCl. Vancomycin-HCl (53 mg) was
dissolved in 6.7 ml of coupling buffer (100 mM NaHCO3, 500 mM NaCl, pH 8) and
subsequently mixed with the sepharose matrix. The suspension was rotated end-over-end
for 2 hrs at room temperature, after which the beads were extensively washed with coupling
buffer (250 ml) to remove the excess of vancomycin ligand. To react any possibly
remaining active succinimide groups, the beads were suspended in 100 mM Tris-HCl, pH
8.0 (10 ml) and incubated for 1 hr at room temperature under gentle agitation. The beads
were washed with 50 mM sodium acetate, 500 mM NaCl, pH 4, and 50 mM Tris-HCl, 500
mM NaCl, pH 8, in three alternating cycles of 50 ml each. Finally, beads were washed once
with 50 ml coupling buffer and stored at 4°C in 10 ml of coupling buffer containing 0.03%
(w/v) NaN3. For purification of peptides crosslinked to Lipid II-ASA, the tryptic digest was
applied to ~400 µl settled vancomycin-sepharose and washed ten times with 200 µl PBS pH
7.5. Bound material was eluted ten times with 200 µl 100 mM Tris, 100 mM NaCl, pH 11,
and radioactivity of each fraction was determined by scintillation counting.
MALDI-TOF mass spectrometry analysis
Mass spectrometry was performed on a Voyager DE MALDI-TOF. Affinity-purified
peptides were vacuum-dried, redissolved in water containing 0.1% (v/v) TFA and applied
to a µ-C18 Zip Tip (Millipore). Peptides were washed with 0.1% TFA and eluted onto a
stainless steel MALDI target with 60% (v/v) acetonitrile, 0.1% TFA (v/v) in water,
containing α-cyano-4-hydroxycinnamic acid matrix (10 mg/ml). Sample spots were
subsequently air-dried at room temperature. Measurements were performed in linear
positive- and negative mode, respectively, at an acceleration voltage of 25 kV, 90% grid
voltage and a delay time of 200 ns.
Results and Discussion
To obtain site-specific information on the binding of Lipid II to E. coli PBP1b, Lipid II
radiolabeled at its GlcNAc moiety and derivatized at its lysine residue with a photo-reactive
azidosalicylic acid (ASA) group ([14C]Lipid II-ASA, Fig. 1A) was prepared as described in
87
A
B
14 C
Pi Pi M
2
8
G
A
E
K
A
A
Figure 1. (A) Schematic structure of Lipid II derivatized at its lysine residue with an azidosalicylic
acid photoreactive crosslinking group and radiolabeled at the GlcNAc moiety ([14C]Lipid II-ASA).
(B) TLC analysis of purified [14C]Lipid II-ASA detected by iodine staining (lane 1). Lipid II
reference standard is in lane 2.
Materials and methods and obtained in high purity as shown by TLC analysis (Fig. 1B).
The bifunctional PBP1b was chosen as it is the major peptidoglycan synthase in E. coli
with activity during both cell growth and division. It has been shown previously that dimers
of PBP1b effectively catalyze the polymerization of Lipid II in vitro [9], allowing it to
serve as a model for other transglycosylases. To test if [14C]Lipid II-ASA is also used as a
substrate for transglycosylation, the radioactive precursor was incubated with PBP1b that
was present at a concentration favoring dimerization [9] for 1 hour at 37 °C. The reaction
mixture was separated by TLC and radioactivity was detected by phosphor imaging (Fig.
2). Lipid II-ASA runs at an Rf ~0.8, as was shown in a control experiment (not shown) and
is in agreement with previous observations for Lipid II [21]. The radioactive signal at the
origin of the TLC corresponds to polymerization products [21]. High amount of
radioactivity was detected at intermediate Rf-values, indicating the presence of reaction
products consisting of a low number of monomeric subunits (Fig. 2). This could be a
consequence of the presence of the photo-reactive group, which prevents transpeptidation
and thereby the formation of higher-order structures. Nonetheless, these data show that
Lipid II-ASA is utilized by PBP1b as a substrate for transglycosylation. Next, crosslinking
of Lipid II to PBP1b was monitored by SDS-PAGE and autoradiography. Polymerization
88
Front
Lipid II-ASA
Origin
Figure 2. Phosphor image of PBP1b catalyzed reaction products of [14C]Lipid II-ASA separated by
TLC.
was allowed to proceed before crosslinking was activated by UV illumination. The reaction
mixture was separated by SDS-PAGE followed by either exposure to X-ray film or staining
with Coomassie Blue. The corresponding autoradiogram showed a band at low molecular
weight, most likely non-crosslinked Lipid II-ASA, and a diffuse zone of radioactivity at
high molecular weight (Fig. 3, lane 1). PBP1b was shown by Coomassie staining to run at
~90 kDa (Fig. 3, lane 5), somewhat lower than the radioactive signal. This indicated that
polymerization products of Lipid II-ASA with heterogeneous lengths were covalently
bound to the protein. To obtain homogeneous crosslink products for subsequent analysis,
the reaction mixture was subjected to digestion with lysozyme. Autoradiogram showed that
the majority of the radioactive zone had narrowed into a band at the molecular weight of
PBP1b, and the amount of radioactive compounds with low molecular weights appeared to
be significantly increased (Fig. 3, lane 2). This indicates that the long glycan strands have
for a large part been digested by lysozyme resulting in a single disaccharide-pentapeptide
unit and/or possibly very short units attached to the protein.
When crosslinking was performed following incubation of Lipid II-ASA and PBP1b in the
presence of the transglycosylation inhibitor moenomycin, one single radioactive band was
observed with a molecular weight similar to the protein (Fig. 3, lane 3). This suggests that
Lipid II-ASA is still able to bind to the enzyme when moenomycin inhibits the formation of
89
peptidoglycan strands, which is in agreement with the presence of two substrate binding
sites.
Figure 3. Autoradiogram (top) of SDS-PAGE analysis of [14C]Lipid II-ASA crosslinking to PBP1b.
Crosslinking was performed in the absence of moenomycin (lane 1), with additional lysozyme
digestion following UV illumination (lane 2), and in the presence of moenomycin (lane 3). SDSPAGE with Coomassie staining (bottom) of PBP1b (lane 5). Lanes 4 and 6 are molecular weight
markers. Bold arrow at high molecular weight in lane 1: transglycosylation products crosslinked to
PBP1b. Arrow at low molecular weight in lane 1: non-reacted [14C]Lipid II-ASA. Arrow in lane 3:
crosslink product with similar molecular weight to PBP1b.
In order to identify the Lipid II crosslinking sites, PBP1b crosslinked to Lipid II in the
presence or absence of moenomycin, was digested into peptides using trypsin. To simplify
subsequent analysis, the long hydrophobic polyprenyl tail was cleaved off by boiling at
mildly acidic pH. The digest was analyzed by HPLC with dual detection of UV and
radioactivity. In the absence of moenomycin, a multitude of radioactive peaks were
observed (data not shown). This indicates that during the process of polymerization,
crosslink labels at various positions in the growing strand are located in close proximity of
the enzyme, resulting in many different crosslinked peptide products after trypsin digestion.
However, when the crosslinking experiment was performed in the presence of
moenomycin, the radioactive HPLC profile strikingly showed only one major peak after
~81 minutes (Fig. 4). Apparently, when moenomycin is bound to the active site, Lipid II
90
crosslinks at one major location within the enzyme, presumably its specific binding site.
The small peak at ~56 minutes is from residual Lipid II headgroup as was determined in
control experiments (data not shown). The low amount of detected radioactivity was found
to be due to limited detector sensitivity (data not shown).
40
35
25
A214
radioactivity (cpm)
30
20
15
10
5
0
0
10
20
30
40
50
60
70
80
90
100
time (min)
Figure 4. C18 HPLC analysis of the tryptic digest of PBP1b crosslinked to [14C]Lipid II-ASA in the
presence of moenomycin. Gray trace: UV absorption at 214 nm, Black trace: Radioactivity after
subtraction of background noise (10 cpm).
To identify the sequence of peptide(s) covalently bound to the Lipid II headgroup, digests
were subjected to affinity purification using a vancomycin-sepharose column. Vancomycin
binds with high affinity to the D-Ala-D-Ala moiety of Lipid II enabling the specific
purification of crosslink products. Monitoring of radioactivity during purification showed
that the majority was retained onto the column and eluted off with 100 mM Tris pH 11 (fig.
5). Subsequently, the entire experiment was repeated using non-radioactive Lipid II-ASA
and the vancomycin-purified fraction was analyzed by mass spectrometry. MALDI-TOF
analysis in linear positive ion mode revealed that multiple peptides were present.
Unfortunately, attempts at PSD fragmentation analysis were unsuccessful due to poor
91
200
180
160
140
cpm
120
100
80
60
40
20
e9
e1
0
e1
1
e1
2
e1
3
e1
4
e7
e8
w
1
pH FT
7.
5
w
2
w
3
w
4
w
5
w
6
w
7
w
8
w
9
w
10
e1 w1
1
e2 pH
e3 pH 9
p
e4 H 9
pH 10
e5
pH 10
11
e6
0
Figure 5. Radioactive profile, as determined by liquid scintillation counting, of the affinity
purification of crosslink products using vancomycin-sepharose. FT: flow-through, w: wash, e: elution.
signal intensity in reflectron mode. In order to identify which of the mass signals, if not all,
corresponded to crosslink products, experimentally measured masses were compared with
theoretical masses obtained from a virtual trypsin digestion of E. coli PBP1b
(http://www.expasy.ch/tools/peptide-mass.html).
Furthermore,
the
presence
of
the
phosphate in the Lipid II headgroup of crosslink products was exploited. Phosphorylated
peptides can easily be recognized by comparing spectra recorded in both positive and
negative ion mode, as their low ionization efficiency in positive ion mode results in a
drastic increase of signal intensity in negative ion mode [22]. Two peaks displayed these
characteristics, observed at m/z 3947 and 5358; [M-H]- (fig. 6). However, the mass
resolution in linear mode was relatively low resulting in low mass accuracy. After
subtraction of the molecular weight of the attached Lipid II part (1179 Da), the resulting
mass was compared with a theoretical tryptic digest allowing maximal two missed
cleavages. Peptides with masses in closest approximation to the calculated masses consisted
of amino acid residues His236-Arg260 (2759.4 Da) and His236-Lys274 (4214.2 Da)
containing one and two missed cleavage sites, respectively. The residues are located within
92
Voyager Spec #1[B P= 2302
. 9, 1863
8]
833
833.3
100
A
90
80
70
% Intensity
60
50
40
30
20
10
0
3300
3840
4380
4920
5460
6000
Mass (m/z)
Voyager Spec #
1[ BP =2310.4, 43102]
100
1241.7
1241
90
B
80
70
% Intensity
60
50
40
30
20
10
0
3300
3840
4380
4920
5460
6000
Mass (m/z)
m/z
Figure 6. MALDI-TOF mass spectrometry analysis of purified crosslink products. Spectra were
recorded in linear negative ion mode (A), and linear positive ion mode (B). Indicated peaks show
drastically increased signal intensity in negative ion mode. Triangle: m/z 3947, Star: m/z 5358.
the transglycosylase domain and contain the residues His240 and Gly242, which are highly
conserved among mono- and bifunctional transglycosylases (Fig. 7). It is possible that the
presence of the crosslinked Lipid II sterically hinders access of trypsin, explaining the
missed cleavages. Also, both crosslink products are in large part identical and could
therefore have similar chromatographic properties, possibly causing overlap of radioactive
signals during HPLC separation. In the crystal structure of E. coli PBP1b in complex with
moenomycin, a loop region consisting of residues 249-267 was absent (Figure 8)[13]. This
possibly indicates that the loop is highly flexible, which could be important for proper
binding of the substrate. Interestingly, in a recent study on the active site of S. aureus
93
217 PRSFGPDLLVDTLLATEDRHFYEHDGISLYSIGRAVLANLTAGRTVQGASTLTQ
271 QLVKNLFLSSERSYWRKANEAYMALIMDARYSKDRILELYMNEVYLGQSGDNEI
325 RGFPLASLYYFGRPVEELSLDQQALLVGMVKGASIYNPWRNPKLALERRNLVLR
379 LLQQQQIIDQELYDMLSARP
Figure 7. Amino acid sequence of the transglycosylase domain of E. coli PBP1b. Residues highly
conserved among transglycosylases are bold and underlined.
Figure 8. Transglycosylase domain of E. coli PBP1b in complex with moenomycin. Amino acid
residues H236-V248 from the putative crosslink site are in orange with side chains, residues 249-267
were absent in the crystal structure. Moenomycin is in dark blue.
94
monofunctional transglycosylase in complex with moenomycin, an ordered phosphate ion
was found in a binding pocket containing the same conserved histidine and glycine,
separated from the moenomycin binding site by the jaw region [16]. The corresponding
sequence in E. coli PBP1b consists of residues His240-Leu245 and could therefore be part
of the Lipid II binding site. In conclusion, we have shown that our photo-crosslinking assay
using radiolabeled Lipid II-ASA is a promising tool to study the interaction between Lipid
II and transglycosylases. Peptide sequence determination is required to confirm the binding
location after which mutational studies could provide more detailed information on
functionally important residues. Considering the essential role of PBP1b in peptidoglycan
synthesis and the high degree of sequence conservation among transglycosylase domains of
class A PBPs and monofunctional transglycosylases, this will pave the way for the
development of new broad-range antibacterial compounds targeting the transglycosylases.
References
1.
2.
3.
4.
5.
6.
7.
8.
Sauvage, E., et al., The penicillin-binding proteins: structure and role in
peptidoglycan biosynthesis. FEMS Microbiol Rev, 2008. 32(2): p. 234-58.
Rojo, F., et al., Analysis of the different molecular forms of penicillin-binding
protein 1B in Escherichia coli ponB mutants lysogenized with specialized
transducing lambda (ponB+) bacteriophages. Eur J Biochem, 1984. 144(3): p.
571-6.
Nakagawa, J. and M. Matsuhashi, Molecular divergence of a major peptidoglycan
synthetase with transglycosylase-transpeptidase activities in Escherichia coli --penicillin-binding protein 1Bs. Biochem Biophys Res Commun, 1982. 105(4): p.
1546-53.
Pepper, E.D., M.J. Farrell, and S.E. Finkel, Role of penicillin-binding protein 1b in
competitive stationary-phase survival of Escherichia coli. FEMS Microbiol Lett,
2006. 263(1): p. 61-7.
Chalut, C., et al., Differential responses of Escherichia coli cells expressing
cytoplasmic domain mutants of penicillin-binding protein 1b after impairment of
penicillin-binding proteins 1a and 3. J Bacteriol, 2001. 183(1): p. 200-6.
Zijderveld, C.A., M.E. Aarsman, and N. Nanninga, Differences between inner
membrane and peptidoglycan-associated PBP1B dimers of Escherichia coli. J
Bacteriol, 1995. 177(7): p. 1860-3.
Zijderveld, C.A., et al., Penicillin-binding protein 1B of Escherichia coli exists in
dimeric forms. J Bacteriol, 1991. 173(18): p. 5740-6.
Chalut, C., M.H. Remy, and J.M. Masson, Disulfide bridges are not involved in
penicillin-binding protein 1b dimerization in Escherichia coli. J Bacteriol, 1999.
181(9): p. 2970-2.
95
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
Bertsche, U., et al., In vitro murein peptidoglycan synthesis by dimers of the
bifunctional transglycosylase-transpeptidase PBP1B from Escherichia coli. J Biol
Chem, 2005. 280(45): p. 38096-101.
Lovering, A.L., et al., Structural insight into the transglycosylation step of
bacterial cell-wall biosynthesis. Science, 2007. 315(5817): p. 1402-5.
Perlstein, D.L., et al., The direction of glycan chain elongation by peptidoglycan
glycosyltransferases. J Am Chem Soc, 2007. 129(42): p. 12674-5.
Terrak, M., et al., Importance of the conserved residues in the peptidoglycan
glycosyltransferase module of the class A penicillin-binding protein 1b of
Escherichia coli. J Biol Chem, 2008. 283(42): p. 28464-70.
Sung, M.T., et al., Crystal structure of the membrane-bound bifunctional
transglycosylase PBP1b from Escherichia coli. Proc Natl Acad Sci U S A, 2009.
106(22): p. 8824-9.
Lovering, A.L., M. Gretes, and N.C. Strynadka, Structural details of the
glycosyltransferase step of peptidoglycan assembly. Curr Opin Struct Biol, 2008.
18(5): p. 534-43.
Yuan, Y., et al., Structural analysis of the contacts anchoring moenomycin to
peptidoglycan glycosyltransferases and implications for antibiotic design. ACS
Chem Biol, 2008. 3(7): p. 429-36.
Heaslet, H., et al., Characterization of the active site of S. aureus monofunctional
glycosyltransferase (Mtg) by site-directed mutation and structural analysis of the
protein complexed with moenomycin. J Struct Biol, 2009. 167(2): p. 129-35.
Kohlrausch, U. and J.V. Holtje, One-step purification procedure for UDP-Nacetylmuramyl-peptide murein precursors from Bacillus cereus. FEMS Microbiol
Lett, 1991. 62(2-3): p. 253-7.
Rouser, G., S. Fkeischer, and A. Yamamoto, Two dimensional then layer
chromatographic separation of polar lipids and determination of phospholipids by
phosphorus analysis of spots. Lipids, 1970. 5(5): p. 494-6.
Breukink, E., et al., Lipid II is an intrinsic component of the pore induced by nisin
in bacterial membranes. J Biol Chem, 2003. 278(22): p. 19898-903.
van den Brink-van der Laan, E., et al., Membrane interaction of the
glycosyltransferase MurG: a special role for cardiolipin. J Bacteriol, 2003.
185(13): p. 3773-9.
Anderson, J.S., et al., Lipid-Phosphoacetylmuramyl-Pentapeptide and LipidPhosphodisaccharide-Pentapeptide:
Presumed
Membrane
Transport
Intermediates in Cell Wall Synthesis. Proc Natl Acad Sci U S A, 1965. 53: p. 8819.
Janek, K., et al., Phosphopeptide analysis by positive and negative ion matrixassisted laser desorption/ionization mass spectrometry. Rapid Commun Mass
Spectrom, 2001. 15(17): p. 1593-9.
96
Chapter 5
Mechanism of action of the
transglycosylation inhibitor 5b
Abstract
With the rapid emergence of multidrug-resistant bacteria, there is an urgent need for new
antibacterial drugs. The transglycosylation step in peptidoglycan biosynthesis is an
excellent target for the development of new antibiotics. It involves the polymerization of
Lipid II forming the linear glycan chains of nascent peptidoglycan and is catalyzed by
transglycosylases. Moenomycin and its derivatives are the only known direct
transglycosylase enzyme inhibitors. Compounds that inhibit transglycosylation by binding
to the Lipid II substrate, such as vancomycin and teicoplanin, have been very important
clinical drugs for the past decades. From a structure-based virtual screening for compounds
inhibiting transglycosylase enzymes, compound 5b (2-(1-(3,4-dichlorobenzyl)-2-methyl-5(methylthio)-1H-indol-3-yl)ethanamine) emerged as a potential candidate. 5b had been
shown to inhibit polymerization of Lipid II by transglycosylases in vitro and kills Grampositive bacteria with MIC values in the micromolar range. In this study, we investigated
whether the mechanism by which 5b inhibits transglycosylation involves direct binding to
the enzyme or an interaction with the substrate Lipid II. It was established that 5b has a
high membrane affinity and is able to dissipate the membrane potential of bacterial
cytoplasmic membranes. Moreover, it was shown that 5b effectively inhibits
carboxyfluorescein leakage induced by nisin-Lipid II pore formation in model membranes
and that this effect was diminished in the presence of a molar excess of
undecaprenylpyrophosphate. This points to a pyrophosphate-mediated interaction of 5b
with Lipid II, which can explain the inhibition of transglycosylation.
Introduction
The cell wall (peptidoglycan) biosynthesis pathway is highly conserved in all bacteria and
presents many targets for antibiotics. The final stages of peptidoglycan biosynthesis involve
the polymerization of the Lipid II precursor by the penicillin-binding proteins (PBPs). The
PBPs belong to either class A or class B, with enzymes from both classes exerting
transpeptidase activity and class A PBPs (bifunctional transglycosylases) containing an
additional transglycosylase (TG) domain responsible for catalyzing the elongation of
glycan chains [1]. Monofunctional transglycosylases have also been found in several
98
bacteria, however their precise function is not well understood [2, 3]. Transglycosylases are
potentially very attractive targets for the design of new antibiotics. Bacterial
transglycosylase inhibitors can be divided into two classes: compounds that achieve
inhibition by binding directly to the active site of the enzyme and those that bind to the
substrate Lipid II, thereby sterically preventing the action of transglycosylases [4, 5]. The
only
comprehensively
studied
broad-range
transglycosylase-binding
inhibitor
is
moenomycin, which was discovered some forty years ago. Moenomycin is extremely
potent against Gram-positive bacteria, but poor pharmacokinetic properties have prevented
its clinical use. Substrate binders include the clinically important glycopeptide antibiotics
(vancomycin, teicoplanin) and the lantibiotics (nisin) [6]. For a long time, attempts to
analyze bacterial transglycosylases were hindered by complexities, such as obtaining
workable quantities of the substrate Lipid II. The large-scale emergence of resistance
against antibiotics has renewed interest in transglycosylases. Recent studies, including the
elucidation of several TG crystal structures, have provided great insight into the mechanism
of transglycosylation and have paved the way for the comprehensive search for novel
inhibitors [7-9]. Using the crystal structure of Staphylococcus aureus PBP2, a structurebased virtual screening for transglycosylase inhibitors was performed, resulting in a set of
promising leads, which were subsequently tested in vitro for inhibitory activity against
several mono- and bifunctional transglycosylases [10]. One compound, (2-(1-(3,4dichlorobenzyl)-2-methyl-5-(methylthio)-1H-indol-3-yl)ethanamine)
(Fig.
1),
further
referred to as 5b, was found to effectively inhibit transglycosylation and possesses
antibacterial activity [10]. In this study, it was investigated whether 5b inhibits
transglycosylation through competitive enzyme inhibition or possibly via complex
formation with the transglycosylase substrate Lipid II. It is shown that 5b has a high
membrane affinity and dissipates the membrane potential of bacterial membranes. NisinLipid II pore formation in model membranes was revealed to be inhibited by 5b.
Undecaprenylpyrophosphate was shown to compete with Lipid II for 5b. This indicates an
interaction of 5b with the pyrophosphate moiety of Lipid II, which, thus, could explain the
mechanism for inhibition of transglycosylation.
99
Figure
1.
Structure
of
5b
(2-(1-(3,4-dichlorobenzyl)-2-methyl-5-(methylthio)-1H-indol-3-
yl)ethanamine.
Materials and methods
Chemicals
1,2-Dioleoyl-sn-glycero-3-phosphocholine (DOPC), 1,2-dioleoyl-sn-glycero-3-[phosphorac-(1-glycerol)]
(DOPG), 1,2-dioleoyl-sn-glycero-3-[phospho-rac-(3-lysyl(1-glycerol))]
(lysyl-DOPG) and 1,2-dioleoyl-3-trimethylammonium-propane (DOTAP) were purchased
from Avanti Polar Lipids Inc. Nisin A was produced, isolated, and purified as described
[11]. E. coli PBP1b, S. aureus MtgA and 5b were kind gifts from A. Derouaux, University
of Liege. Staphylococcus simulans and Micrococcus flavus were grown in LB broth at 37
ºC and 30 ºC, respectively, with aeration. Membrane potential measurements were
performed using the fluorescent dye 3,3‘-diethylthiodicarbocyanine iodide (DiSC2(5)) from
Molecular Probes Inc. Lipid I and Lipid II were synthesized and purified as described
elsewhere [12]. Undecaprenylphosphate and undecaprenylpyrophosphate were obtained by
phosphorylation of undecaprenol [13] that was isolated from Laurus nobilis as described
[14].
Preparation of large unilamellar vesicles (LUVs)
100
Large unilamellar vesicles were prepared essentially as described [15]. Desired amounts of
lipid solutions in chloroform were mixed and evaporated under a gentle stream of nitrogen.
The lipid film was subsequently dried for 20 min under vacuum. The film was hydrated by
the addition of the buffer of choice under mechanical agitation and submitted to ten freezethaw cycles using liquid nitrogen and a water bath. The lipid suspension was then extruded
ten times through a polycarbonate membrane filter with a pore size of 200 nm (Whatman
International) [16]. The final phospholipid concentration was determined by phosphate
analysis according to [17].
Binding assay
DOPC with or without 1 mol% Lipid II vesicles were prepared as described above in 10
mM MES-KOH, 15 mM K2SO4 at pH 7. Vesicles (1 mM lipid Pi) were incubated with 5
µM and 20 µM 5b, respectively, for 15 min at RT. The mixture was centrifuged in a TLA
120.2 rotor using a Beckman Ultracentrifuge (TL-100) for 1.5 h at 100 krpm and 20 °C.
The amount of 5b before centrifugation and in the supernatant and pellet was determined by
fluorescence on a Cary Eclipse fluorescence spectrophotometer (Varian Inc.) using
excitation and emission wavelengths of 250 nm and 350 nm, respectively.
Carboxyfluorescein leakage assay in large unilamellar vesicles
Carboxyfluorescein-loaded LUVs were prepared in 50 mM MES-KOH, 100 mM K2SO4,
pH 6.5 (K+-buffer) as described above additionally containing 50 mM CF. Following the
extrusion step, the vesicle suspension was applied to Sephadex G-50 spin columns
equilibrated with K+-buffer to remove free CF. The final phospholipid concentration was
determined by phosphate analysis according to Rouser et al. Vesicles were resuspended in
K+-buffer to a concentration of 25 µM and 5b was added at the desired concentration 1 min
prior to the addition of nisin A (50 nM, final conc.). High-salt experiments were conducted
in K+-buffer supplemented with 0.5M NaCl. The nisin-induced CF leakage from the
vesicles was monitored with excitation and emission wavelengths set at 430 nm and 513
nm, respectively. Triton X-100 was added 1 min after the addition of nisin, to a final
concentration of 0,2% (w/v) to fully disrupt the lipid vesicles and the corresponding
fluorescence was taken as 100% leakage.
Membrane permeabilization assay
101
The cytoplasmic membrane depolarization activity of 5b was determined with the
membrane potential-sensitive dye DiSC2(5) (16) using S. simulans and M. flavus. Bacterial
cells in the mid-logarithmic phase were centrifuged, washed in 5 mM HEPES (pH 7.8), and
resuspended in the same buffer to an optical density at 600 nm of 0.05 in a 1-cm cuvette. A
stock solution of DiSC2(5) was added to a final concentration of 0.4 µM and quenching was
allowed to occur at room temperature for ~1 min. The desired concentration of 5b was
added and changes in fluorescence due to the disruption of the membrane potential gradient
across the cytoplasmic membrane were continuously recorded with a Cary Eclipse
spectrofluorometer at an excitation and emission wavelengths of 622 and 670 nm,
respectively.
Results
Inhibition of transglycosylase activity by 5b can be achieved either through direct binding
to the active site of the enzyme or via binding to the substrate Lipid II. To study a possible
direct interaction between 5b and transglycosylases, ITC measurements were conducted
(data not shown). The monofunctional transglycosylase MtgA from S. aureus and the
bifunctional PBP1b from E. coli were titrated with 5b. Despite the controlled presence of
detergent and DMSO in the buffers, major heat effects were observed during the titration
that did not appear to be the result of 5b-protein interactions. This severely complicated
data analysis. Nevertheless, at all concentrations tested, no apparent binding of 5b to either
enzyme
could
be
observed.
Therefore,
we
investigated
whether
5b
inhibits
transglycosylation by binding to the common transglycosylase substrate Lipid II. ITC was
performed using 5b and DOPC model membranes, either with or without 2 mol% Lipid II
(data not shown). The titration curve of the vesicles containing Lipid II was not
significantly different from that of the pure DOPC vesicles. However, the obtained binding
curves suggested binding of 5b to the lipid vesicles, which could obscure a possible
interaction with Lipid II.
To further study the binding of 5b to Lipid II containing LUVs, a binding assay was
performed using vesicles with the same lipid composition as in the ITC experiments. After
incubation with 5b for 15 minutes, vesicles were spun down by ultracentrifugation. Using
the characteristic intrinsic fluorescence of 5b, its respective percentage in the pellet and
supernatant was determined. 5b was present exclusively in the pellet in both LII-containing
102
and pure DOPC vesicles (Fig. 2). Control experiments showed that in the absence of
vesicles, 5b alone did not pellet during ultracentrifugation (Fig. 2). This shows that 5b has a
very high affinity for the lipid bilayer, regardless of whether Lipid II is present.
100
90
80
% of total 5b
70
60
supernatant
50
pellet
40
30
20
10
0
5b only
+ DOPC
+ DOPC/Lipid II
Figure 2. Membrane binding assay using 5b. The interaction of 5b with Lipid II containing DOPC
LUVs was studied using ultracentrifugation. The assay was performed in 10 mM MES/Tris, 15 mM
K2SO4, pH 7.0 with 1 mM DOPC vesicles containing 1% Lipid II (mol/mol) and 20 µM 5b.
Excitation wavelength was at 250 nm and emission was monitored from 320 to 400 nm. Percentage of
5b in the supernatant and pellet was determined by comparing the maximal value of fluorescence
emission intensity.
To test for a possible specific interaction between 5b and Lipid II, a competition assay was
set up with the lantibiotic nisin, which uses Lipid II for its pore-forming activity [12, 18,
19]. A similar approach had previously been used with vancomycin in the study that
determined the binding of nisin to Lipid II [20]. The addition of nisin to
carboxyfluorescein-loaded vesicles containing Lipid II causes leakage of the fluorescent
dye from the vesicles, due to Lipid II-dependent pore formation by nisin. This leakage
results in an increase of fluorescence as self-quenching of the dye is reduced. The leakage
103
can be expressed as a percentage of the maximum fluorescence intensity after complete
disruption of the membranes by the addition of Triton X-100. When 50 nM nisin was added
to 25 µM DOPC vesicles containing 0,1 mol% Lipid II, a leakage of ~75% was observed.
In the presence of 2,5 µM 5b, CF leakage was significantly decreased by ~36%. Moreover,
increasing concentrations of 5b resulted in higher reductions and almost complete inhibition
of leakage at 15 µM (Fig. 3). These results could suggest that 5b binds to Lipid II, thereby
preventing an interaction of the latter with nisin.
100
% leakage reduction
80
60
40
20
0
0
2
4
6
8
10
12
14
16
5b (µM)
Figure 3. 5b interference with nisin pore formation in DOPC LUVs containing 0,1 mol % Lipid II. 5b
was added prior to nisin (50 nM final conc.) and the carboxyfluorescein leakage was monitored. The
5b-induced reduction of leakage is expressed as a percentage of the leakage observed in the absence
of 5b relative to the total release of carboxyfluorescein after addition of 0.2% (w/v) Triton X-100. The
average percentage from two independent experiments are shown, error bars depict the spread
between results.
To gain insight into the mechanism by which 5b inhibits nisin-induced CF leakage and the
possible importance of the lipid composition herein, leakage experiments were
subsequently conducted in LUVs containing anionic phospholipids, composed of
DOPC/DOPG (1:1 mol/mol) and 0.1 mol% Lipid II. In the presence of low concentrations
of 5b, pore-forming activity of nisin was inhibited, as was observed in DOPC/Lipid II
104
vesicles. However, at 5b concentrations of 10 µM and higher, an increase in CF
fluorescence was already observed before the addition of nisin that appeared to be
proportional to the added concentration of 5b. This indicated that the functional integrity of
the membrane was disrupted by the presence of 5b at high concentrations. To test whether
the presence of Lipid II was important for this effect, 5b was added at the same
concentrations to DOPC/DOPG (1:1) vesicles without Lipid II. Comparable leakage of the
fluorescent dye was observed following the addition of 5b (Fig. 4), revealing that this was a
general membrane effect independent of Lipid II. Since DOPC vesicles with or without
Lipid II displayed negligible leakage in the presence of 5b (Fig. 4), this indicates that 5b is
able to disturb the membrane integrity in an anionic lipid-dependent way. To investigate if
5b exerts similar effects on bacterial membranes, the depolarization effect of 5b on the
plasma membrane in intact cells of Staphylococcus simulans and Micrococcus flavus was
determined using the membrane-potential sensitive fluorophore DiSC2(5). This dye inserts
100
% leakage
80
PC/PG (1:1)
PC
60
40
20
0
0
10
20
30
40
5b (µM)
Figure 4. Disruption of membrane integrity by 5b in DOPC/DOPG (1:1 mol/mol) vesicles.
Carboxyfluorescein leakage, expressed as a percentage of the total release after addition of 0.2%
(w/v) Triton X-100, following the addition of various concentrations of 5b to 25 µM DOPC and
DOPC/DOPG (1:1 mol/mol) vesicles, respectively. The results from a representative experiment are
shown.
105
into the cytoplasmic membrane in the presence of a membrane potential resulting in selfquenching of fluorescence. The addition of 5b to a final concentration of 1 µM caused a
major release of the dye from the membrane in M. flavus (Fig. 5) showing that the
membrane potential was dissipated upon binding of 5b. S. simulans also showed significant
membrane depolarization at a somewhat higher 5b concentration. Nisin was more effective
in dissipating the membrane potential in M. flavus (Fig. 5), which is in relative agreement
with the difference in MIC values for both compounds (~5 nM nisin for M. flavus versus
~10 µM 5b for most Gram-positive strains tested (Derouaux)).
500
fluorescence (a.u.)
400
300
200
M. flavus Nisin (100 nM)
S. simulans 5b (5 µM)
M. flavus 5b (1 µM)
100
0
0,0
0,2
0,4
0,6
0,8
1,0
1,2
time (min)
Figure 5. Cytoplasmic membrane depolarization of Micrococcus flavus and Staphylococcus simulans
by 5b, assessed using the membrane potential sensitive fluorophore DiSC2(5). Effect of nisin in
Micrococcus flavus is shown for comparison. Dye release was monitored at excitation and emission
wavelengths of 622 and 670 nm, respectively.
As nisin is a cationic peptide, another possible mechanism by which 5b could inhibit nisin
activity could be to repel nisin from the membrane by creating a positive surface charge.
Therefore, the effect of a positively charged membrane on nisin activity was tested. CFloaded DOPC/Lipid II vesicles were prepared containing 10, 20 or 30% of the cationic
lipids DOTAP and lysyl-DOPG, respectively. Following the addition of nisin, a
106
considerable reduction of CF leakage was observed with increasing concentrations of either
cationic lipid (Fig. 6). This shows that inhibition of nisin by 5b could in principle be due to
electrostatic repulsion at the membrane surface.
100
90
% leakage reduction
80
70
60
50
40
30
lysyl-DOPG
20
DOTAP
10
0
0
10
20
30
40
mol% cationic lipid
Figure 6. Influence of cationic lipids on nisin activity. Carboxyfluorescein leakage was determined
following the addition of 50 nM nisin in DOPC/LII vesicles containing respectively 0, 10, 20 or 30
mol% of the cationic lipids DOTAP or lysyl-PG. Leakage reduction is expressed as in Figure 3.
If 5b would specifically interact with Lipid II, structurally related compounds could
compete for binding. In order test this possibility, the inhibition effect of 5b on nisin-Lipid
II pore formation was analyzed in the presence of the peptidoglycan biosynthetic
intermediates undecaprenylphosphate (11-P) and undecaprenylpyrophosphate (11-PP),
respectively. Nisin activity in the presence and absence of 5b was analyzed in CF-loaded
DOPC vesicles containing 0,1 mol% Lipid II and 2 mol% of either 11-P, 11-PP or the
anionic lipid DOPG. Strikingly, in the 11-PP containing vesicles CF-release induced by
nisin was not significantly reduced in the presence of 2,5 µM 5b (Fig. 7). In comparison, 5b
was still able to cause significant inhibition of nisin activity in the presence of 11-P or
DOPG, albeit somewhat less profound than in their absence (Fig. 7). This indicates that the
excess of 11-PP competes with Lipid II for binding 5b and that the pyrophosphate moiety is
107
important for the interaction. 5b activity was also studied in CF-loaded DOPC vesicles
containing Lipid I instead of Lipid II. Leakage due to pore formation of nisin with Lipid I
was diminished similarly by the presence of 5b as with Lipid II (data not shown),
suggesting that the GlcNAc residue of Lipid II is not an essential requirement for the
interaction.
% leakage reduction by 2,5 µM 5b
50
45
40
35
30
25
20
15
10
5
0
DOPC +0,1% LII DOPC +0,1% LII DOPC +0,1% LII DOPC +0,1% LII
+2% 11-PP
+2% 11-P
+2% DOPG
Figure 7. Competition for 5b binding. Inhibition by 5b of nisin-induced CF leakage from DOPC
vesicles containing undecaprenylpyrophosphate (11-PP), undecaprenylphosphate (11-P) and DOPG,
respectively, in 20-fold molar excess over Lipid II was analyzed. The 5b-induced reduction of
leakage is depicted as a percentage of the total leakage observed in the absence of 5b. The averages of
two independent experiments are shown with the error bars indicating the spread between values.
Discussion
In this study, we investigated the mechanism of action of the antibacterial agent 5b. This
small amphipathic molecule had previously been shown in vitro to inhibit the
transglycosylation reaction that polymerizes Lipid II into peptidoglycan strands. The
activity of transglycosylases can be inhibited by 5b either through occupying the active site
of the enzyme or by complexation of Lipid II preventing its utilization as a substrate. ITC
measurements did not provide any evidence for a direct interaction of 5b with E. coli
108
PBP1b and S. aureus MtgA and results from titration experiments with vesicles containing
Lipid II were also inconclusive. Binding assay revealed that 5b has a high membrane
affinity, which is a likely explanation for the inconclusive data obtained from ITC analysis
using Lipid-II containing vesicles. In relation to this, several lines of evidence implied that
5b possesses an additional mode of antibacterial action. First, after our binding assay
showed that 5b has a very high membrane affinity, carboxyfluorescein leakage experiments
revealed that above a certain concentration threshold 5b induced disruption of the
functional integrity of anionic PC/PG membranes. The depolarization of membranes in
bacteria, as monitored with the membrane potential-sensitive probe DiSC2(5), confirmed
this membrane targeting mode of action. Furthermore, it was previously reported that
inhibition of the transglycosylase reaction in Gram-positive bacteria does not result in the
rapid bactericidal effects exerted by 5b [10, 21]. This indicates that targeting the bacterial
membrane is the predominant mechanism of antibacterial action of 5b. We hypothesize that
the hydrophobic character of the molecule allows for effective insertion into the membrane,
whereby the amine moiety presumably interacts with the polar lipid headgroups. The
distinct action towards anionic and zwitterionic membranes is most likely explained by
differences in orientation and penetration of 5b with respect to the membrane, comparable
to what is observed with antimicrobial peptides. Recently, a novel antibacterial agent, XF73, was shown to specifically target the cytoplasmic membrane of Gram-positive bacteria
including MRSA, rapidly killing cells without the emergence of resistance after many
repeated passages [22]. Although its exact mechanism of action is still unknown, it is
currently in clinical development demonstrating the huge potential of membrane targeting
drugs.
A
striking observation
was the
effective
inhibition by 5b
of
nisin-induced
carboxyfluorescein leakage from Lipid II containing vesicles. Inhibition by 5b was more
efficient than vancomycin in the same assay [20]. In principle, 5b inhibition of pore
formation by nisin with Lipid II could occur in two ways. The first possibility is through
complex formation with Lipid II. As it was shown that 5b has a very high membrane
affinity, another mechanism could be by altering the surface charge of the membrane
resulting in the electrostatic repulsion of nisin. Although antimicrobial compounds that act
at the cell surface and/or target the bacterial membrane are very diverse when it comes to
size, structure and mode of action, the majority carries a net cationic charge, which
facilitates binding to the anionic bacterial cell surface through electrostatic attraction and
109
thereby enhances their efficacy. Changing the lipid composition and thereby the
electrostatic properties of its membrane is an important defence mechanism of bacteria
against these compounds. The presence of a relatively high amount of the cationic lipid
lysyl-PG, which is formed by the attachment of lysine to the headgroup of the major
anionic membrane lipid PG, is suggested to play a vital role in the pathogenicity and multidrug resistance of S. aureus. The amine group of 5b, which was previously shown to be
essential for activity [10], renders the compound positively charged. Binding of 5b could
therefore change the electrostatic properties of the vesicle membrane in that way
disfavoring binding of the highly cationic nisin, thus explaining the inhibition effect. The
addition of increasing amounts of either lysyl-PG or DOTAP in the PC/Lipid II vesicles
significantly reduced nisin activity, which can thus most likely be attributed to the rising
electrostatic repulsion between the polypeptide and the membrane surface. Several
observations make it unlikely, however, that 5b inhibits nisin activity in a similar manner.
Firstly, in CF-loaded PC/PG (1:1) vesicles containing Lipid II, low concentrations of 5b
significantly reduced nisin-induced leakage. The presence of 50% anionic lipids makes it
unattainable for 5b to generate a cationic membrane surface charge equivalent to that of the
vesicles containing DOTAP or lysyl-PG. Furthermore, a 20-fold molar excess of
undecaprenylpyrophosphate over Lipid II in DOPC vesicles drastically affected 5b in its
ability to inhibit nisin activity. Thus, the additional presence of undecaprenylpyrophosphate
allowed nisin to gain more access to Lipid II, which indicates that 5b reduces nisin-induced
pore formation by binding to Lipid II and that undecaprenylpyrophosphate competes with
Lipid II for interaction with 5b. The pyrophosphate moiety appears to be essential for the
interaction with 5b as undecaprenylphosphate and DOPG did not significantly compete
with Lipid II for 5b. As nisin also binds the pyrophosphate of Lipid II, this could explain
the effective inhibition by 5b of nisin-Lipid II pore assembly. Nisin is also known to bind
undecaprenylpyrophosphate without forming pores [23, 24], which could be a reason for
the somewhat lower leakage compared to DOPC/LII vesicles in the absence of 5b. The
amine residue of 5b had been shown to be essential for its transglycosylation inhibiting
activity, suggesting its crucial involvement in the interaction with the pyrophosphate group
of Lipid II.
Finally, we still cannot rule out the possibility that 5b interacts with TG enzymes, but the
observations described here indicate that binding to the substrate Lipid II comprises the
mechanism by which 5b inhibits transglycosylation.
110
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
Sauvage, E., et al., The penicillin-binding proteins: structure and role in
peptidoglycan biosynthesis. FEMS Microbiol Rev, 2008. 32(2): p. 234-58.
van Heijenoort, J., Formation of the glycan chains in the synthesis of bacterial
peptidoglycan. Glycobiology, 2001. 11(3): p. 25R-36R.
Derouaux, A., et al., The monofunctional glycosyltransferase of Escherichia coli
localizes to the cell division site and interacts with penicillin-binding protein 3,
FtsW, and FtsN. J Bacteriol, 2008. 190(5): p. 1831-4.
Halliday, J., et al., Targeting the forgotten transglycosylases. Biochem Pharmacol,
2006. 71(7): p. 957-67.
Ostash, B. and S. Walker, Bacterial transglycosylase inhibitors. Curr Opin Chem
Biol, 2005. 9(5): p. 459-66.
Breukink, E. and B. de Kruijff, Lipid II as a target for antibiotics. Nat Rev Drug
Discov, 2006. 5(4): p. 321-32.
Sung, M.T., et al., Crystal structure of the membrane-bound bifunctional
transglycosylase PBP1b from Escherichia coli. Proc Natl Acad Sci U S A, 2009.
106(22): p. 8824-9.
Lovering, A.L., et al., Structural insight into the transglycosylation step of
bacterial cell-wall biosynthesis. Science, 2007. 315(5817): p. 1402-5.
Yuan, Y., et al., Crystal structure of a peptidoglycan glycosyltransferase suggests
a model for processive glycan chain synthesis. Proc Natl Acad Sci U S A, 2007.
104(13): p. 5348-53.
Derouaux, A.e.a., Discovery of new small molecules antibacterial targeting
peptidoglycan glycosyltransferases. unpublished.
Kuipers, O.P., et al., Engineering dehydrated amino acid residues in the
antimicrobial peptide nisin. J Biol Chem, 1992. 267(34): p. 24340-6.
Breukink, E., et al., Lipid II is an intrinsic component of the pore induced by nisin
in bacterial membranes. J Biol Chem, 2003. 278(22): p. 19898-903.
Danilov, L.L., et al., Polyprenyl phosphates: synthesis and structure-activity
relationship for a biosynthetic system of Salmonella anatum O-specific
polysaccharide. Chem Phys Lipids, 1989. 51(3-4): p. 191-203.
Swiezewska, E., et al., The search for plant polyprenols. Acta Biochim Pol, 1994.
41(3): p. 221-60.
Breukink, E., et al., The C-terminal region of nisin is responsible for the initial
interaction of nisin with the target membrane. Biochemistry, 1997. 36(23): p.
6968-6976.
Hope, M.J., et al., Production of Large Unilamellar Vesicles by a Rapid Extrusion
Procedure - Characterization of Size Distribution, Trapped Volume and Ability to
Maintain a Membrane-Potential. Biochimica Et Biophysica Acta, 1985. 812(1): p.
55-65.
Rouser, G., S. Fkeischer, and A. Yamamoto, Two dimensional then layer
chromatographic separation of polar lipids and determination of phospholipids by
phosphorus analysis of spots. Lipids, 1970. 5(5): p. 494-6.
111
18.
19.
20.
21.
22.
23.
24.
van Heusden, H.E., B. de Kruijff, and E. Breukink, Lipid II induces a
transmembrane orientation of the pore-forming peptide lantibiotic nisin.
Biochemistry, 2002. 41(40): p. 12171-8.
Wiedemann, I., et al., Specific binding of nisin to the peptidoglycan precursor lipid
II combines pore formation and inhibition of cell wall biosynthesis for potent
antibiotic activity. J Biol Chem, 2001. 276(3): p. 1772-9.
Breukink, E., et al., Use of the cell wall precursor lipid II by a pore-forming
peptide antibiotic. Science, 1999. 286(5448): p. 2361-4.
Baizman, E.R., et al., Antibacterial activity of synthetic analogues based on the
disaccharide structure of moenomycin, an inhibitor of bacterial transglycosylase.
Microbiology, 2000. 146 Pt 12: p. 3129-40.
Ooi, N., et al., XF-73, a novel antistaphylococcal membrane-active agent with
rapid bactericidal activity. J Antimicrob Chemother, 2009. 64(4): p. 735-40.
Hsu, S.T., et al., The nisin-lipid II complex reveals a pyrophosphate cage that
provides a blueprint for novel antibiotics. Nat Struct Mol Biol, 2004. 11(10): p.
963-7.
Bonev, B.B., et al., Targeting extracellular pyrophosphates underpins the high
selectivity of nisin. FASEB J, 2004. 18(15): p. 1862-9.
112
Chapter 6
Summarizing discussion
This thesis describes the application of novel analysis approaches aimed at gaining new
insights into the physiology of the bacterial cell wall. Labeled peptidoglycan precursors
were utilized as a tool for in vivo cell wall metabolism and proteomics studies.
Furthermore, the role of Lipid II as a target for transglycosylase inhibition is highlighted.
In vivo labeling of the cell wall using derivatized peptidoglycan precursors
A new approach to incorporate reporter groups into the bacterial cell wall of E. coli is
described in Chapter 2. It involves supplying cells with the cell wall precursor murein
tripeptide (AeK) fluorescently labeled with NBD. It is shown that this precursor is taken up
by the cell and incorporated into the cell wall by means of the cell wall recycling pathway.
The choice for the lysine form rather than diaminopimelic acid (DAP), which is common to
E. coli, was to considerably simplify the chemical synthesis of the NBD-labeled tripeptide,
since amine- and carboxylic acid protected DAP was not available. Neither the use of
lysine, nor the presence of the fluorescent label was expected to affect incorporation
efficiency to a large extent. It was previously shown that lysine-containing tripeptide,
common to Gram-positive bacteria, also serves as a substrate for E. coli Mpl in vitro,
although with decreased catalytic efficiency compared to the tripeptide containing DAP [1].
Upon addition to UDP-MurNAc by Mpl, tripeptides containing either lysine or DAP serve
about equally well as substrates for MurF, the enzyme responsible for the subsequent
attachment of the D-Ala- D-Ala moiety [1]. Furthermore, supplying E. coli inner membrane
vesicles with NBD-labeled cell wall precursors was shown to result in the formation of
NBD-Lipid II [2]. The presence of the label is unlikely to affect transglycosylation, whereas
only transpeptidation is partly impaired as the peptide can only act as a donor in peptide
cross bridge formation.
Using the novel approach, the major amidase activity of AmiC during cell division could be
visualized. This was determined through analysis of the labeling pattern of septal
peptidoglycan in wild-type cells and amidase mutants. The FtsZ dependent hydrolase
activity during the preseptal stage of peptidoglycan synthesis, observed in cell division
mutants, is especially intriguing. Although it is clear that specific enzymes localize to the
division site in a temporal order, the exact nature of events surrounding initiation of cell
division remain unclear. There are several implications that specific hydrolase activity
could be involved in the differentiation of the site where ensuing division is to take place. It
is well established that the composition and/or three-dimensional structure of peptidoglycan
114
in a zone at mid-cell is altered, in a thus far unknown way, so that it becomes metabolically
stable or ‘inert’. Two important players in this process are the cell division protein FtsZ and
the carboxypeptidase PBP5. Morphology defects are associated with misplacement and/or
accumulation of inert peptidoglycan, with the deletion of pentapeptide regulation by PBP5
playing the most influential part and deletion of additional hydrolases visibly enhancing the
defects [5, 6]. FtsZ is known to be essential for the initiation of cell division as it
polymerizes at mid-cell into the Z-ring, which forms the main scaffold for the assembly of
the entire complex of division proteins. Recently, a new role for FtsZ in guiding
peptidoglycan synthesis during cell elongation was established [3, 4]. In E. coli strains
lacking PBP5, which have a much higher pentapeptide content in the cell wall (>30%) as
compared to hardly detectable amounts in wild-type strains (<0.5%), depletion of FtsZ
resulted in a reduced precursor incorporation into the side walls near the poles, together
with a significant increase in the pentapeptide levels [7]. In an unrelated study using PBP5
mutants, AmiA and AmiB, but not AmiC, were shown to cleave pentapeptides from the
peptidoglycan strands [8]. As pentapeptides can function as both a donor and an acceptor in
peptide crosslinking, FtsZ-directed synthesis probably produces more peptide crosslinks by
utilizing pentapeptide-containing muropeptides, which could in turn be the determining
factor for peptidoglycan to become inert. In order for polar peptidoglycan synthesis to
commence and cells to constrict, the inert portion of the lateral wall at mid-cell must first be
cleaved to a certain degree by peptidoglycan hydrolases. This could point to a specific role
for AmiA and/or AmiB in removing pentapeptides prior to the initiation of polar
peptidoglycan synthesis. Endopeptidases cleaving peptide crosslinks are also likely to be
involved. Components of the division-specific protein complex that are recruited at a later
timepoint, including AmiC, are likely to be responsible for maintaining the correct balance
between continued insertion of new material and formation of inert polar peptidoglycan
during the continued course of cell division (Fig. 1). In an interesting observation, the
putative amidase AmiA of Helicobacter pylori was implicated to be essential for the
morphological transition from bacillary to coccoid form [9].
115
PBP5
Tetra/tripeptides
Recyclable PG
Complete divisome
Septal PG
FtsZ
Inert ‘preseptal’ PG
Pentapeptides
Hydrolases (AmiA/B?)
Figure 1. Schematic representation of the initiation and continuation of cell division. Important
enzymes and peptidoglycan characteristics for each stage are indicated.
Our in vivo cell wall labeling approach was expanded to a proteomics format in chapter 3.
By using a murein tripeptide analogue equipped with a photo-activatable crosslinker,
proteins interacting with the cell wall and its precursors were covalently captured by UV
illumination. The additional presence of an alkyne tag in the derivatized peptide allowed for
straightforward in vitro detection of crosslinked proteins via click chemistry using a
fluorescent azide tag. Purification of crosslink products was initially attempted using a twostep procedure involving biotinylation by means of click chemistry followed by affinity
capture onto neutravidin beads. Non-specifically purified proteins interfered with protein
identification, which prompted us to develop a purification approach involving magnetic
azide beads. Combining the high specificity of click chemistry with the low non-specific
binding associated with the magnetic beads, it resulted in a higher recovery and more
reliable identifications. One limitation for the general application of this purification
method is the inability to release the intact molecule from the beads for subsequent
analysis. The insertion of a cleavable linker between the azide functionality and the bead
would therefore expand its applicability. No enzymes known to be directly involved in
peptidoglycan biosynthesis were identified, which could be explained by the transient
nature of their interactions with cell wall precursors. Crosslink labels residing in the mature
116
cell wall have more chance of encountering a stable interaction partner, as evidenced by the
identification of proteins such as major outer membrane lipoprotein (Lpp), outer membrane
porins and several secretion system proteins.
In the labeling studies described here, E. coli cells were grown under the selective pressure
of peptidoglycan precursors. One possible consequence could be a shift in the intracellular
balance between peptidoglycan precursors, e.g. relative accumulation of tripeptides. This
balance has been shown to play a crucial role in the differential expression of selected
genes associated with antibiotic resistance mechanisms. The β-lactamase enzymes are
responsible for resistance against β-lactam antibiotics, as they hydrolyze the essential
lactam ring. The expression of β-lactamases is induced by the accumulation of
anhydromuropeptides upon exposure to the antibiotics. These peptidoglycan degradation
products compete with UDP-MurNAc-pentapeptide for a binding site on the transcriptional
regulator AmpR, causing activation of ampC that encodes the AmpC β-lactamase [10, 11].
In another example, increased levels of precursors containing a D-Ala-D-Lac sequence
were found in vancomycin-resistant strains. Vancomycin normally binds to the terminal DAla-D-Ala sequence of Lipid II, thereby blocking its incorporation. This modification was
shown to decrease the binding affinity of vancomycin by a factor over 1000 [12]. The
synthesis and biochemical modification of the peptidoglycan precursors in resistant strains
is mediated by enzymes encoded by van operons. Expression of the van genes is activated
by the VanS/VanR two-component system in response to extracellular glycopeptide
antibiotic [13]. The identity of the VanS effector ligand is not completely clear, but the
accumulation of Lipid II is believed to play a direct role in the induction mechanism of
vancomycin resistance [14]. In our labeling approach, cell wall synthesis is not blocked and
cell growth was shown not to be greatly affected, making it improbable that major changes
occur in the precursor pool, nevertheless, it remains tempting to investigate whether any
changes are induced in gene expression of E. coli.
The role of Lipid II in targeting peptidoglycan transglycosylases
Photo-crosslinking was used for the purpose of characterizing the interaction of the major
bifunctional transglycosylase-transpeptidase PBP1b with its natural substrate Lipid II
(Chapter 4). For this purpose, radiolabeled Lipid II derivatized with a photocrosslinker was
prepared. UDP-MurNAc-pentapeptide was labeled at its lysine residue with N117
hydroxysuccinimidyl-4-azidosalicylic acid (NHS-ASA) and used to synthesize Lipid IASA, which was subsequently enzymatically converted to radiolabeled-Lipid II-ASA using
UDP-[14C]GlcNAc. PBP1b was shown to utilize the photoreactive substrate for
transglycosylation reactions and polymerization products could be crosslinked to the
enzyme upon UV exposure. In the presence of the transglycosylase inhibitor moenomycin,
Lipid II was found to crosslink to one predominant position in the enzyme. The capacity of
vancomycin to bind Lipid II was utilized to affinity purify peptide crosslink products
derived from the enzyme. Peptide sequences could not be directly identified by mass
spectrometry analysis due to poor signal intensities, however, crosslink products could be
detected using a method to detect phosphopeptides. Calculated peptide masses pointed to a
crosslink site within the transglycosylase domain. Interestingly, the majority of this putative
crosslink site was absent from the crystal structure of the enzyme. In possible agreement
with this, ligand binding sites in proteins characteristically occur in multiple structural
conformations in both bound and unbound state [15, 16].
The mode of action of the transglycosylase inhibitor 5b was investigated in Chapter 5.
Compound 5b inhibits the activity of transglycosylases from different bacterial strains and
is bactericidal for Gram positive bacteria. A specific interaction between enzyme and
inhibitor could not be detected, while several observations pointed to an interaction with
Lipid II. 5b effectively inhibited the pore-forming activity of the lantibiotic nisin in
complex with Lipid II in model membrane vesicles loaded with the fluorescent dye
carboxyfluorescein. Furthermore, the presence of an excess of undecaprenylpyrophosphate
over Lipid II drastically reduced this effect, indicating a competition between interaction
partners of 5b. In light of the striking structural complexity of known Lipid II-binding
transglycosylase inhibiting compounds, such as vancomycin, ramoplanin and nisin, the
intriguing question arises how 5b accomplishes the same feat with its relatively simple
structure.
The lipoglycodepsipeptide ramoplanin was shown to form a complex with Lipid II, most
likely in a dimeric form, aggregating into insoluble fibrils upon binding [17, 18]. The
pyrophosphate moiety of Lipid II was shown to be required for the interaction, with the
positively charged ornithine residues of ramoplanin proposedly playing an essential role
[18, 19]. The lantibiotic nisin kills bacteria by forming highly stable membrane pores in
complex with Lipid II [20]. Nisin also establishes interaction with Lipid II via the
118
pyrophosphate group [21, 22]. The first two lanthionine rings form a cage-like structure
around the pyrophosphate moiety, with the backbone amides stabilizing the interaction
through hydrogen bonds [22]. Resistance against the cationic nisin is mainly associated
with the alteration of the net negative cell surface charge in a way to prevent it from
reaching Lipid II. A decrease in anionicity of teichoic acids and the membrane
phospholipid composition have been reported in nisin-resistant strains [23, 24]. Reduction
of membrane affinity inhibits the activity of nisin, as was shown in an extreme example in
our leakage experiments conducted in overall positively charged membrane vesicles. No
resistance has yet been found against the positively charged ramoplanin, which is most
likely due to the presence of its lipophilic moiety that directs targeting to the bacterial
membrane, regardless of less favorable electrostatic conditions. The high membrane
affinity of 5b, which was established in the binding assay, is likely initially driven by
hydrophobicity. The amine moiety of 5b has been shown to be essential for
transglycosylase inhibition and is therefore likely to be involved in the putative interaction
with the pyrophosphate group of Lipid II.
Recently, two compounds, DMPI and CDFI, with somewhat similar structures to 5b were
shown to inhibit peptidoglycan synthesis and restore susceptibility of MRSA to β-lactams
[25]. The target of these compounds was shown to be the gene product of SAV1754 in S.
aureus, the Gram-positive homolog of E. coli MurJ [26], a flippase proposed to be
responsible for the translocation of Lipid II across the cytoplasmic membrane. Inactivation
of SAV1754, as well as a number of other genes encoding for cell wall biosynthesis
proteins, such as PBP2 and FemA, was shown to potentiate β-lactams activity against
MRSA. It could therefore be possible that DMPI and CDFI bind to Lipid II, thereby
inhibiting activity of these enzymes and contributing to the sensitisation of MRSA to βlactams.
Concluding remarks
The importance of Lipid II as target for antibiotics is well established. In times of
increasing bacterial resistance against currently used antibiotics, the full characterization of
the Lipid II interaction with peptidoglycan transglycosylases will provide the necessary
information for design and discovery of novel antibacterial drugs. One such compound is
5b, for which further studies are needed to shed more light on the structural details of its
putative interaction with Lipid II. Finally, labeled murein tripeptide precursors were shown
119
to have great potential as tools for bacterial cell wall studies in Escherichia coli and
possibly other cell wall recycling bacteria. Straight-forward synthesis and the ability to
employ a variety of labels give the method a wide applicability. One could for example
imagine a combination with fluorescently labeled proteins to study interactions with the cell
wall and the inhibition thereof using visualization by fluorescence resonance energy
transfer. Preliminary studies in this regard using the NBD-labeled tripeptide and
tetramethylrhodamine labeled vancomycin in E. coli triple amidase mutants have already
shown promising results. Further optimization should therefore make the in vivo cell wall
labeling approach a useful asset to the field of peptidoglycan analysis.
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
Baum, E.Z., et al., Utility of muropeptide ligase for identification of inhibitors of
the cell wall biosynthesis enzyme MurF. Antimicrob Agents Chemother, 2006.
50(1): p. 230-6.
van Dam, V., et al., Transmembrane transport of peptidoglycan precursors across
model and bacterial membranes. Mol Microbiol, 2007. 64(4): p. 1105-14.
Aaron, M., et al., The tubulin homologue FtsZ contributes to cell elongation by
guiding cell wall precursor synthesis in Caulobacter crescentus. Mol Microbiol,
2007. 64(4): p. 938-52.
Varma, A., M.A. de Pedro, and K.D. Young, FtsZ directs a second mode of
peptidoglycan synthesis in Escherichia coli. J Bacteriol, 2007. 189(15): p. 5692704.
Young, K.D., Bacterial shape. Mol Microbiol, 2003. 49(3): p. 571-80.
Varma, A. and K.D. Young, FtsZ collaborates with penicillin binding proteins to
generate bacterial cell shape in Escherichia coli. J Bacteriol, 2004. 186(20): p.
6768-74.
Varma, A., M. de Pedro, and K.D. Young, FtsZ directs a second mode of
peptidoglycan synthesis in Escherichia coli. J Bacteriol, 2007.
Priyadarshini, R., D.L. Popham, and K.D. Young, Daughter cell separation by
penicillin-binding proteins and peptidoglycan amidases in Escherichia coli. J
Bacteriol, 2006. 188(15): p. 5345-55.
Chaput, C., et al., Role of AmiA in the morphological transition of Helicobacter
pylori and in immune escape. PLoS Pathog, 2006. 2(9): p. e97.
Jacoby, G.A., AmpC beta-lactamases. Clin Microbiol Rev, 2009. 22(1): p. 161-82,
Table of Contents.
Uehara, T. and J.T. Park, Role of the murein precursor UDP-N-acetylmuramyl-LAla-gamma-D-Glu-meso-diaminopimelic acid-D-Ala-D-Ala in repression of betalactamase induction in cell division mutants. J Bacteriol, 2002. 184(15): p. 4233-9.
Bugg, T.D., et al., Molecular basis for vancomycin resistance in Enterococcus
faecium BM4147: biosynthesis of a depsipeptide peptidoglycan precursor by
vancomycin resistance proteins VanH and VanA. Biochemistry, 1991. 30(43): p.
10408-15.
120
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
Hong, H.J., M.I. Hutchings, and M.J. Buttner, Vancomycin resistance VanS/VanR
two-component systems. Adv Exp Med Biol, 2008. 631: p. 200-13.
Baptista, M., et al., Single-cell analysis of glycopeptide resistance gene expression
in teicoplanin-resistant mutants of a VanB-type Enterococcus faecalis. Mol
Microbiol, 1999. 32(1): p. 17-28.
Mobley, D.L. and K.A. Dill, Binding of small-molecule ligands to proteins: "what
you see" is not always "what you get". Structure, 2009. 17(4): p. 489-98.
Ma, B., et al., Multiple diverse ligands binding at a single protein site: a matter of
pre-existing populations. Protein Sci, 2002. 11(2): p. 184-97.
Hu, Y., et al., Ramoplanin inhibits bacterial transglycosylases by binding as a
dimer to lipid II. J Am Chem Soc, 2003. 125(29): p. 8736-7.
Hamburger, J.B., et al., A crystal structure of a dimer of the antibiotic ramoplanin
illustrates membrane positioning and a potential Lipid II docking interface. Proc
Natl Acad Sci U S A, 2009. 106(33): p. 13759-64.
Cudic, P., et al., Complexation of peptidoglycan intermediates by the
lipoglycodepsipeptide antibiotic ramoplanin: minimal structural requirements for
intermolecular complexation and fibril formation. Proc Natl Acad Sci U S A,
2002. 99(11): p. 7384-9.
Breukink, E., et al., Use of the cell wall precursor lipid II by a pore-forming
peptide antibiotic. Science, 1999. 286(5448): p. 2361-4.
Bonev, B.B., et al., Targeting extracellular pyrophosphates underpins the high
selectivity of nisin. FASEB J, 2004. 18(15): p. 1862-9.
Hsu, S.T., et al., The nisin-lipid II complex reveals a pyrophosphate cage that
provides a blueprint for novel antibiotics. Nat Struct Mol Biol, 2004. 11(10): p.
963-7.
Verheul, A., et al., Modifications of membrane phospholipid composition in nisinresistant Listeria monocytogenes Scott A. Appl Environ Microbiol, 1997. 63(9): p.
3451-7.
Crandall, A.D. and T.J. Montville, Nisin resistance in Listeria monocytogenes
ATCC 700302 is a complex phenotype. Appl Environ Microbiol, 1998. 64(1): p.
231-7.
Huber, J., et al., Chemical genetic identification of peptidoglycan inhibitors
potentiating carbapenem activity against methicillin-resistant Staphylococcus
aureus. Chem Biol, 2009. 16(8): p. 837-48.
Ruiz, N., Bioinformatics identification of MurJ (MviN) as the peptidoglycan lipid
II flippase in Escherichia coli. Proc Natl Acad Sci U S A, 2008. 105(40): p.
15553-7.
Pinho, M.G., H. de Lencastre, and A. Tomasz, An acquired and a native
penicillin-binding protein cooperate in building the cell wall of drug-resistant
staphylococci. Proc Natl Acad Sci U S A, 2001. 98(19): p. 10886-91.
121
Nederlandse samenvatting
Bacteriën zijn eencellige organismen die een belangrijke rol spelen in het leven van de
mens. In positieve zin zijn bacteriën in het menselijk lichaam verantwoordelijk voor onder
andere het goed functioneren van het digestieve systeem en de productie van belangrijke
stoffen zoals vitamine K. In negatieve zin zijn bacteriën vooral bekend als verwekkers van
vele ziektes, waaronder hersenvliesontsteking, tuberculose en de recent in het nieuws
gekomen Q-koorts. De ontdekking van penicilline, een kleine eeuw geleden, leidde tot de
ontwikkeling van een breed scala aan natuurlijke en synthetische stoffen die bacteriën
doden of hun groei remmen; de antibiotica. De afgelopen decennia is er een sterke toename
van bacteriesoorten die resistent zijn tegen een grote groep antibiotica, zoals de beruchte
‘ziekenhuisbacterie’ MRSA (methicilline resistente Staphylococcus aureus). Het is daarom
van cruciaal belang nieuwe strategieën en doelwitten te vinden om de groei van pathogene
bacteriën te remmen. Een van de primaire doelwitten van antibiotica is de bacteriële
celwand. Deze weerstaat de hoge interne osmotische druk en is daarom essentieel voor de
overleving van de cel. Bovendien bevindt de celwand zich aan de buitenkant van de cel
waardoor het relatief goed bereikbaar is. Bekende antibiotica als penicilline en
vancomycine verzwakken de celwand door de opbouw ervan tijdens verschillende stadia te
blokkeren. Er bestaan nog veel onduidelijkheden rondom de biochemische processen
waarbij de celwand betrokken is. Een beter begrip van deze processen is nodig voor het
vinden van nieuwe doelwitten voor antibiotica, waarvoor additionele analysemethoden
moeten worden ontwikkeld.
Het voornaamste bestanddeel van de celwand is peptidoglycaan. Het is een polymeer
opgebouwd uit lineaire ketens van alternerende aminosuikers, N-acetylglucosamine
(GlcNAc) en N-acetylmuraminezuur (MurNAc), welke onderling worden verbonden door
korte peptidebruggen. Dit resulteert in een rigide driedimensionaal netwerk. De biosynthese
van peptidoglycaan begint binnen in de cel, in het cytoplasma, waar de aminosuikers
gesynthetiseerd worden en vijf aminozuren sequentieel aan UDP geactiveerde MurNAc
worden gekoppeld. Meestal bestaat de sequentie van het pentapeptide uit L-alanine-Dglutamaat-L-lysine/meso-diaminopimelinezuur-D-alanine-D-alanine.
UDP-MurNAc-
pentapeptide wordt vervolgens aan het membraan gekoppeld via de lipide drager
undecaprenylfosfaat resulterende in Lipide I. De additie van GlcNAc leidt tot de vorming
123
van Lipide II dat vervolgens over het membraan wordt getransporteerd. Aan de buitenkant
van de cel wordt de bouwsteen ingebouwd in de bestaande celwand door de penicilline
bindende eiwitten (PBPs). Deze enzymen katalyseren de elongatie van de suikerketens
(transglycosylatie), en de formatie van de peptide bruggen (transpeptidatie). Het
vrijgekomen undecaprenylpyrofosfaat wordt gedefosforyleerd en terug naar de binnenkant
van het membraan getransporteerd voor een nieuwe synthese cyclus.
Voor het invoegen van nieuwe stukken peptidoglycaan moeten verbindingen verbroken
worden en tijdens de celdeling moeten van de dochtercellen de twee nieuwe peptidoglycaan
polen van elkaar losgemaakt worden. De enzymen verantwoordelijk voor deze afbraak zijn
de peptidoglycaan hydrolases, welke worden onderverdeeld in verschillende groepen
afhankelijk van de posities waar ze de bindingen verbreken. Het veel als modelbacterie
gebruikte Escherichia coli heeft vele hydrolases welke schijnbaar overlappen in functie,
waardoor het moeilijk is om een gedefinieerde functie aan elk individueel enzym toe te
schrijven. De amidases, die het peptide deel in zijn geheel van de suikerketen loskoppelen,
spelen tijdens celdeling ogenschijnlijk de belangrijkste rol. In bepaalde bacteriesoorten,
waaronder E. coli, worden de afbraakproducten van peptidoglycaan op een efficiënte
manier gerecycled. Ze komen via specifieke permeases de cel in en worden uiteindelijk
verwerkt
tot
het
tripeptide
L-alanine-D-glutamaat-meso-diaminopimelinezuur.
Dit
tripeptide wordt hergebruikt voor peptidoglycaan synthese door in zijn geheel aan UDPMurNAc te worden gekoppeld via het enzym murein peptide ligase (Mpl), waarna het
resulterende UDP-MurNAc-tripeptide de biosynthese route vervolgt.
De synthese van peptidoglycaan is nauw gecoördineerd met de bacteriële cel cyclus. Bij
staafvormige bacteriën zoals E. coli wordt tijdens elongatie nieuw peptidoglycaan over de
lengte van de cel ingebouwd en tijdens celdeling vindt de synthese uitsluitend op de
delingsplek plaats. Celdeling wordt geïnitieerd door de polymerizatie van het eiwit FtsZ,
die resulteert in de vorming van een ring-achtig structuur (Z-ring), de localizatie van een
aantal eiwitten en de synthese van een ring van zeer stabiel peptidoglycaan op de
toekomstige
delingsplaats.
Hierna
lokaliseert
een
tweede
groep
eiwitten
die
verantwoordelijk zijn voor de formatie van het septum en de scheiding van de
dochtercellen. Sommige eiwitten betrokken bij peptidoglycaan synthese en afbraak, zoals
PBP3 en de amidase AmiC, behoren tot deze groep. Er bestaan nog veel onduidelijkheden
over de processen die plaatsvinden vlak voor en tijdens de celdeling evenals over de
identiteit en functie van vele van de betrokken eiwitten.
124
Om meer inzicht te krijgen in de processen waarbij de celwand is betrokken, is er in dit
proefschrift een methode ontwikkeld om het peptidoglycaan in levende E. coli cellen te
labelen met specifieke groepen die voor verscheidene doeleinden kunnen dienen. Deze
methode maakt gebruik van gelabelde bouwstenen van de celwand. Deze bouwstenen,
gelabelde varianten van het tripeptide L-alanine-D-glutamaat-L-lysine, worden opgenomen
door de cel en langs metabolische weg ingebouwd door middel van het peptidoglycaan
recycling systeem. Dit wordt in hoofdstuk 2 aangetoond met het tripeptide gelabeld met een
fluorescente NBD groep aan de lysine zijketen (AeK-NBD). Nadat wild-type E. coli cellen
met het fluorescente peptide waren geïncubeerd kon er zowel fluorescent gelabeld Lipide II
als peptidoglycaan uit de cellen worden geïsoleerd. Door de aanwezigheid van de
fluorescente groepen in de celwand, konden bepaalde metabolische processen in de cellen
gevolgd worden met behulp van een confocale fluorescentiemicroscoop. Elongerende
cellen hadden een uniform gelabelde celwand, maar bij cellen die in het delingsproces zaten
werd een aanzienlijke afname van de fluorescentie op de delingsplaats waargenomen. Om
te kunnen onderscheiden of het gebrek aan fluorescentie veroorzaakt werd doordat de
gelabelde bouwstenen niet werden ingebouwd of door afbraak van gelabeld peptidoglycaan,
werd het labelingspatroon in mutanten met meerdere inactieve amidases geanalyseerd.
Hieruit werd duidelijk dat het gelabelde peptide wel in het septale peptidoglycaan wordt
ingebouwd en dat AmiC betrokken is bij de afbraak van gelabeld peptidoglycaan in delende
wild-type cellen. Labelingspatronen in celdelingsmutanten suggereerden dat er een of
meerdere nog te identificeren hydrolases actief zijn tijdens de FtsZ-afhankelijke
overgangsfase van celelongatie naar celdeling.
In hoofdstuk 3 hebben we de labelingsmethode aangepast tot een proteomics aanpak.
Proteomics is de grootschalige identificatie en analyse van eiwitten. Bij het metabolisme
van peptidoglycaan is een groot en complex netwerk van eiwitten betrokken die potentiële
doelwitten vormen voor antibiotica. Daarnaast zijn eiwitten die een interactie met de
celwand hebben belangrijk bij andere processen zoals pathogenese en secretie. De opzet
was om tijdens het inbouwen van het peptide in de celwand van levende cellen, eiwitten die
een interactie hebben met de celwand of de bouwstenen/afbraakproducten daarvan te
crosslinken en vervolgens te identificeren. Hiervoor werd een tripeptide zoals in hoofdstuk
2 gesynthetiseerd met aan de lysine zijketen in plaats van een fluorescent groep een label
bestaande uit een foto-activeerbare crosslinker en een alkyn groep. Tijdens het inbouwen
125
van dit peptide in de celwand van E. coli cellen werd de crosslink groep geactiveerd door
bestraling met UV licht. Hierdoor wordt het zeer reactief en vormt het een covalente
binding met moleculen die zich in de nabijheid bevinden. De crosslink producten konden
specifiek worden gedetecteerd en gezuiverd door in vitro geschikte functionele groepen met
een azide groep aan de alkyn groep te koppelen via de zeer specifieke en efficiënte reactie
tussen deze twee chemische groepen, aangeduid als klik reactie. Voor detectie-doeleinden
werd het fluorescente Alexa488-azide gebruikt. Eiwitten werden gescheiden via gel
electroforese en de gecrosslinkte eiwitten konden specifiek worden gedetecteerd door
fluorescent scannen van de gel. Het zuiveren van de crosslink producten werd aanvankelijk
geprobeerd door het koppelen van een biotine groep, gevolgd door affiniteitszuivering met
neutravidine-bolletjes.
De
identificatie
door
massa
spectrometrie
werd
echter
gecompliceerd door aspecifieke binding van eiwitten aan de bolletjes. Om deze reden werd
een andere zuiveringsstrategie ontwikkeld waarbij direct gebruikt werd gemaakt van de
hoge specificiteit van klik-chemie. De gecrosslinkte eiwitten werden covalent aan
magnetische bolletjes met azide groepen gebonden. Er werd aangetoond dat weinig eiwitten
aspecifiek aan deze magnetische bolletjes bonden en bovendien was het door de covalente
binding mogelijk rigoureuzere wasstappen uit te voeren. Diverse eiwitten werden
geïdentificeerd waarvan bekend is dat ze aan de celwand binden of welke mogelijk een
interactie hebben met de celwand, wat door aanvullend onderzoek geverifieerd zal moeten
worden. Hiermee is het potentieel van deze methode aangetoond, die nieuwe doelwitten
voor antibiotica zou kunnen opleveren.
In hoofdstuk 4 en 5 is het polymerisatieproces van de Lipide II suikergroepen
(transglycosylatie) onder de loep genomen. Tot voor kort was er weinig over dit proces
bekend terwijl het een zeer aantrekkelijk doelwit vormt voor de ontwikkeling van nieuwe
antibacteriële stoffen. Enerzijds zijn stoffen die het proces verstoren door aan het Lipide II
substraat te binden zeer geschikt om te dienen als antibiotica. Vancomycine, wat lange tijd
als laatste redmiddel werd gebruikt, is een voorbeeld van een antibioticum wat via dit
mechanisme werkt. Anderzijds is de enige stof waarvan bekend is dat die transglycosylatie
remt door aan de actieve plaats van transglycosylase enzymen te binden, moenomycine,
zeer effectief in het remmen van de groei van bacteriën. Doordat het niet goed wordt
opgenomen door het menselijk lichaam opgenomen, is moenomycine echter ongeschikt
voor de behandeling van infecties. Het is bekend dat moenomycine transglycosylatie remt
door de bindingsplaats te bezetten voor de groeiende suikerketen, waar nieuwe
126
peptidoglycaan subeenheden afkomstig van Lipide II aan wordt gekoppeld. Over de plaats
waar Lipide II aan het enzym bindt is minder bekend. Voor de ontwikkeling van nieuwe
antibiotica die aan het actieve deel van transglycosylases binden, is het belangrijk
gedetailleerde structurele informatie over deze bindingsplaats te verkrijgen. In hoofdstuk 4
is getracht met behulp van crosslinken de bindingsplek van Lipide II in PBP1b van E. coli
te identificeren. PBP1b is een bifunctionele transglycosylase en wordt gezien als het
belangrijkste en meest actieve peptidoglycaan synthetiserende enzym. Radioactief gelabeld
Lipide II gederivatiseerd met een crosslink groep werd geprepareerd. Het werd aangetoond
dat PBP1b deze analoog van Lipide II gebruikt als substraat voor in vitro transglycosylatie
reacties en dat de reactieproducten aan het enzym gecrosslinkt konden worden door
bestraling met UV licht. Door gecrosslinkte eiwitten met het protease trypsine af te breken
en de resulterende peptiden te analyseren met hoge prestatie vloeistof chromatografie
(HPLC) met radioactieve detectie, kon worden vastgesteld dat wanneer transglycosylatie
geblokkeerd werd door moenomycine, het radioactieve substraat voornamelijk op een
specifieke positie bond. De aan Lipide II gebonden peptiden werden vervolgens gezuiverd
met behulp van geïmmobiliseerde vancomycine. Omdat het niet mogelijk bleek om met
behulp van massa spectrometrie structurele informatie te verkrijgen, werd voor de
identificatie van het peptidesequentie de massa van het crosslink product vergeleken met de
molecuulmassa’s van de theoretisch te verkrijgen peptiden na een virtuele digestie van
PBP1b met het protease trypsine. Hieruit volgde dat Lipide II mogelijk was gebonden aan
een deel van het transglycosylase domein met daarin een aantal geconserveerde residuen.
Dit deel zou derhalve de specifieke Lipide II bindingsplaats van het enzym kunnen
omvatten.
In hoofdstuk 5 is het werkingsmechanisme onderzocht van (2-(1-(3,4-dichlorobenzyl)-2methyl-5-(methylthio)-1H-indol-3-yl)ethanamine), aangeduid als 5b, een stof die de
activiteit remt van transglycosylases. Er werd getest of 5b dit bereikt door een directe
interactie met het enzym aan te gaan of met het Lipide II substraat een complex vormt.
Door middel van isotherme titratie calorimetrie kon er geen directe binding van 5b met
transglycosylases worden geconstateerd. Met behulp van modelmembranen kon worden
vastgesteld dat 5b een hoge membraanaffiniteit heeft en in de aanwezigheid van negatief
geladen lipiden de membraan integriteit aantast. Ook in levende bacteriën werd het
membraanpotentiaal aangetast door 5b. Om een mogelijke interactie tussen 5b en Lipide II
vast te kunnen stellen werd een competitie assay toegepast met nisine, een antibioticum dat
127
een complex met Lipide II vormt wat resulteert in de formatie van poriën in het membraan.
In deze assay werden carboxyfluoresceine bevattende membraan vesikels geprepareerd met
in het membraan een kleine hoeveelheid Lipide II aanwezig. Toevoeging van nisine aan
deze vesikels resulteert door de porievorming van nisine met Lipide II in lekkage van
carboxyfluoresceine, hetgeen met fluorescentiespectroscopie kan worden gemeten. Er kon
worden vastgesteld dat de door nisine geïnduceerde lekkage significant werd gereduceerd
door de aanwezigheid van 5b, wat suggereerde dat 5b de formatie van poriën voorkomt
door aan Lipide II te binden. Wanneer het structureel aan Lipide II gerelateerde
undecaprenylpyrofosfaat in een overmaat ten opzichte van Lipide II in de membranen
aanwezig was, werd de vermindering van lekkage voor een significant deel teniet gedaan.
Vermoedelijk werd dit veroorzaakt door competitie tussen undecaprenylpyrofosfaat en
Lipide II voor een interactie met 5b. Deze resultaten impliceren dat binding aan Lipide II
het mechanisme is waarmee 5b transglycosylatie inhibiteerd, waarbij de pyrofosfaat groep
waarschijnlijk een cruciale rol speelt.
Tot slot, er is in dit proefschrift aangetoond dat gelabelde tripeptide precursors van
peptidoglycaan veel potentieel hebben als gereedschap voor studies aan de bacteriële
celwand. De mogelijkheid om verschillende labels te gebruiken maakt deze methode
toepasbaar voor verschillende doeleinden.
Ook is het belang van het aangrijpen van de interactie tussen Lipide II en transglycosylases
benadrukt. Structurele opheldering van deze interactie is nodig voor de ontwikkeling van
nieuwe antibiotica, wat cruciaal is in deze tijden van snel toenemende resistentie.
128
Dankwoord
Dan nu het gedeelte wat voor velen het meest belangrijke deel van een proefschrift vormt.
Om niemand tekort te doen bedank ik de volgende personen in alfabetische volgorde.
Allereerst Lipid II clan-leader Eefjan. Je was een ideale begeleider, die me een hoop
vrijheid gaf en op de goede momenten bijsprong om de boel weer een beetje op de rails te
krijgen. Je labtafel vormde ook een directe inspiratiebron. Verder ken ik je de laatste tijd
natuurlijk beter als CAEB. Je hebt me wegwijs gemaakt in de boze wereld van het online
pokeren, tot het moment dat ik op m’n eigen benen kon staan. Bedankt voor alles!
De volgende die ik wil bedanken is Ben. Je vermogen om op een heldere manier de
essentiële dingen over te brengen, maakte onze meetings altijd zeer vruchtbaar. Ik zag je
vaak als het wetenschappelijke equivalent van Messi die het op z’n heupen heeft...
Natuurlijk was ik graag je laatste aio geweest, maar ik denk dat met Yvonne de cirkel echt
rond voor je is. Het was een voorrecht om met je gewerkt te mogen hebben, geniet van je
pensioen!
Over Ytje gesproken, we hebben bijna hetzelfde traject doorlopen de afgelopen 5 jaar. Je
was altijd het zonnetje van het lab (misschien op die paar maanden voor de deadline na...)
en een hele fijne collega. Laten we er een mooi feest van maken. Vinnie, met jou ook een
hoop gedeeld (artikel, student, bureau, bed (zakelijk), wat nog meer... ellende). Je
anekdotes, humor en hoge octaven waren onmisbaar voor de sfeer en daarnaast heb je me
vaak genoeg met van alles geholpen. Wat rijmt er ook al weer op Jetta? Een minder fijne
collega was Michal (misschien omdat je er minder vaak was...). Onze gesprekken over de
belangrijkste zaken in het leven (niet basketbal) en niet te vergeten een wmv-tje op z’n tijd,
hielpen me zo weer een week door. Ik moet meteen denken aan je grote vriend, Mob-1
Raja, aka Mobster aka Herr Moby alias du verdammtes... What’s the address of your igloo
up there on the North Pole, I can send you a copy. Hope all is well with your family.
Emmalina, you brought a lot of spark to the lab. We had tons of fun, best of luck with your
PhD. I would also like to express my gratitude to the Wayne Gretzky and Mark Messier of
Medicinal Chemistry, Chris and Nathaniel (you can fight over who is who). Both of you
provided me with essential ingredients, taught me a great deal about organic synthesis and
were fun to be around.
129
Tanneke, onze collaboratie liep als een rode draad door de jaren heen. Ik bedank je voor je
zeer belangrijke aandeel in de totstandkoming van hoofdstuk 2. Laten we hopen dat we er
een mooi eind aan kunnen breien. Waldemar, although my car almost broke in two, I very
much enjoyed my short stay in Tubingen. Thank you for your useful suggestions over the
years.
In dit dankwoord mogen mijn studenten natuurlijk ook niet ontbreken. Lakshmi, your 2
minutes and 13 seconds to perform a 20-slide presentation still stands in the Guinness book
of records. Don’t forget your promise to mention those who helped you get started, when
you receive the Noble prize. It was great having you around. Anika, achter ‘ideale student’
in het woordenboek zou een fotootje van jou kunnen staan. Ook bij jouw oefensessies voor
je presentatie moesten we het eerste kwartier altijd reserveren voor slappe lachbuien.
Jammer dat je er maar voor zo’n korte tijd was, al dacht jij daar misschien zelf anders over
na TLC numero 834. Succes in de toekomst!
Martijn alias MCK42, onze diepgaande wetenschappelijke (=poker) gesprekken gingen
nergens over (behalve poker) maar soms moesten we ze met tegenzin afbreken anders
duurde de pauze te lang. Je dronken tilt-acties aan de play money tafel zal ik niet snel
vergeten. Je partners in crime kan ik uiteraard niet vergeten. Mandy, je bent de enige die de
afgelopen vijf jaar meer meiden geproduceerd heeft dan ikzelf. Daarnaast bracht je ook nog
eens het nodige leven in de brouwerij. Blijf aan je Spaans werken. Marlies, het was altijd
fijn als er tenminste nog één iemand aan de koffietafel kon lachen om mijn misplaatste
grappen. Als ik zo mooi kon dichten als je toenmalige protegé had ik het hier zeker gedaan.
Succes met de volgende stap! Het kloppend hart van de afdeling, Irene, bedankt voor al je
hulp de afgelopen jaren. Ik vergeef je voor al die strenge mailtjes, vooral omdat ze
(meestal) ook terecht waren. Diaan, ik weet dat je af en toe stapel werd van mijn... hoe zou
ik het noemen... laksheid, om maar te zwijgen over m’n gewauwel over Italiaanse films. Nu
ik het er toch over heb, ‘I pugni in tasca’ is wel een aanrader. Jacques, do you know
something about computers? We all know about the legendary Sinterklaas poem you wrote,
but your recitation of the one you received (first sentence smooth Dutch, the rest
incomprehensible English) might even be higher on the list of most hilarious moments
during my stay. Thanks for all your help, too. Kamergenoten en Hell’s Angels, Erica en
Tania, er wordt nog steeds met veel lof gesproken over de door ons vlekkeloos
georganiseerde labdag (iedereen naar het toilet van een vriendelijke locale inwoonster van
Odijk). Succes allen met de laatste loodjes. Tami, er is eigenlijk maar één ding wat ik tegen
130
jou kan zeggen: WOIO. Bedankt voor je hulp bij het bedenken van vijftig vogelsoorten in
drie minuten...
Ook de spijkerharde mandekker en de aalgladde linksbuiten uit het glorieteam van FC
Biochemie seizoen ‘77/78, Jan en Dick, wil ik bedanken voor respectievelijk hulp bij mijn
eerste stappen en het draaiende houden van alles op het lab.
Ik besef dat ik op deze manier nog wel even bezig, dus ga ik er maar wat meer vaart achter
zetten. Stafleden Antoinette en Toon, (oud-)collegae Robert (“van mijn part vliegt de hele
boel in de hens”), Pieter, Nicole, Qin (Tami will bring you a copy when she starts her new
job in Shanghai), Robin, Rutger, Thomas, Lucie, Mans, Renke, kamergenoten Remko
(voorheen MrK00, nu MrMeLi), Joost, Paulien, Ed, Jacob, Andrea, Aline, studenten Irene
del O., Roti, ‘schurk van Urk’ Hendrik, Maiku, Jessica, de NIOK voetbalkampioenen 2009
(mede bereikt dankzij vele onbestrafte tackles van Peter S.) en degenen die ik per ongeluk
nog vergeten ben, allen bedankt/thanks/gracias voor jullie geboden hulp en/of babbel.
Ten slotte nog een woordje voor de steun en toeverlaat buiten ‘werk’: familie en vrienden.
Daaf, met betrekking tot de wetenschap hoef ik eigenlijk maar een ding te zeggen: NEJM.
Respect! Ik hoop dat je als paranimf de boel net zo goed organiseert en regelt als ik bij
jouw promotie... Wat betreft het sociale aspect, en dan richt ik me natuurlijk ook naar de
George Best van IJFC, Johan Verweij, kijk ik met plezier terug op en vooruit naar de
nodige verlichtende avondjes. Mozes, je halfjaarlijkse bezoekjes zijn altijd letterlijk en
figuurlijk kort, zo ook dit bedankje. Planken, dank je voor het zo nu en dan delen van je
passie voor chemie.
Pa en ma, bedankt voor jullie steun in alles wat ik deed. Het is misschien niet van het
kaliber ‘Ome Rinus’ of ‘de verloren Air Jordans’, maar ik hoop toch dat jullie ook plezier
aan dit boekje zullen beleven. Ri, we zien elkaar wat minder frequent wat wellicht te maken
heeft met de 20000 km afstand tussen onze woonplaatsen. Leuk dat je er bij kunt zijn, we
hoppen binnenkort ook even langs. Sa en Jasper, het wordt weer tijd voor een coin battle, ik
wil revanche voor m’n smadelijke nederlaag. I nakonets moi tri devushki, moia zhizn, ja
vas ochen ljoebljoe. Voor een vertaling bel of mail Olga Prisekina Translations...
All in.
Nick
131
List of publications
van Dam, V., Olrichs, N. K., Breukink, E., Specific labeling of peptidoglycan precursors as
a tool for bacterial cell wall studies. Chembiochem. 2009 10(4):617-24
Olrichs, N. K., Aarsman, M. E. G., Arnusch, C. J., Vollmer, W., de Kruijff, B., Breukink,
E., den Blaauwen, T. A novel in vivo cell wall labeling approach sheds new light on
peptidoglycan synthesis in Escherichia coli. Submitted.
Derouaux, A., Turk, S., Gobec, S., Olrichs, N. K., Breukink, E., Amoroso, A., Offant, J.,
Vernet, T., Bostock, J., Mariner, K., Chopra, I., Zervosen, A., Joris, B., Frère, J. M.,
Nguyen-Distèche, M. and Terrak M. Discovery of new small molecules antibacterial
targeting peptidoglycan glycosyltransferases. Manuscript in preparation.
132
Curriculum Vitae
Nick Kenji Olrichs was born on January 5th 1980 in Delft, the Netherlands. After
graduating from the Anna van Rijn College in Nieuwegein, he studied Chemistry at Utrecht
University. He did a research project on the purification and identification of N-linked
glycans from glycoproteins at the department of Bio-organic Chemistry under supervision
of dr. Sean Austin and prof. dr. Hans Kamerling. He obtained his masters degree in
September 2003 with honors (‘met genoegen’). Following triumphs in the popular tv game
shows ‘Lucky Letters’ and ‘Weekend Miljonairs’, he traveled to Australia for a sabbatical.
In March 2005, he started his Ph.D. research at the department Biochemistry of Membranes
at Utrecht University under supervision of prof. dr. Ben de Kruijff and dr. Eefjan Breukink,
which resulted in this thesis. In June 2009 he started as a post-doc in the same group.
133
134
Abbreviations
11-P/ 11-PP
AeK
ASA
CF
DAP
DEAE
DIC
DiSC2(5)
DOPC
DOPG
DOTAP
ECL
EDTA
ESI
GlcNAc
HEPES
HPLC
HRP
ITC
Lac
LC-MS/MS
LPS
LUV
lysyl-PG
MALDI-TOF
MES
MS
MudPIT
MurNAc
MWCO
NBD
NMR
PBP
PBS
Pi
RT
SDS-PAGE
TAMRA
TBTA
TCEP
TFA
TG
TLC
TP
Tris
UDP
undecaprenyl phosphate/ undecaprenyl pyrophosphate
L-alanyl-D-glutamyl-L-lysine
azidosalicylic acid
carboxyfluorescein
diaminopimelic acid
diethylaminoethyl
differential interference contrast
3,3‘-diethylthiodicarbocyanine iodide
1,2-dioleoyl-sn-glycero-3-phosphocholine
1,2-dioleoyl-sn-glycero-3-[phospho-rac-(1-glycerol)]
1,2-dioleoyl-3-trimethylammonium-propane
enhanced chemiluminescence
ethylenediaminetetraacetic acid
electrospray ionization
N-acetylglucosamine
4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid
high performance liquid chromatography
horseradish peroxidase
isothermal titration calorimetry
lactate
liquid chromatography-tandem mass spectrometry
lipopolysaccharide
large unilamellar vesicle
1,2-dioleoyl-sn-glycero-3-[phospho-rac-(3-lysyl(1-glycerol))]
matrix-assisted laser desorption/ionization-time of flight
2-(N-morpholino)ethanesulfonic acid
mass spectrometry
multidimensional protein identification technology
N-acetylmuramic acid
molecular weight cut off
N-7-nitro-2, 1, 3-benzoxadiazol-4-yl
nuclear magnetic resonance
penicillin binding protein
phosphate buffered saline
phosphate
room temperature
sodium dodecyl sulfate-polyacrylamide gel electrophoresis
tetramethyl-6-carboxyrhodamine
tris[(1-benzyl-1H-1,2,3-triazol-4-yl)methyl]amine
tris(2-carboxyethyl)phosphine
trifluoroacetic acid
transglycosylase
thin layer chromatography
transpeptidase
trishydroxymethylaminomethane
uridine diphosphate
135
136